InΒ [Β ]:
# This Python 3 environment comes with many helpful analytics libraries installed
# It is defined by the kaggle/python Docker image: https://github.com/kaggle/docker-python
# For example, here's several helpful packages to load
import numpy as np # linear algebra
import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)
# Input data files are available in the read-only "../input/" directory
# For example, running this (by clicking run or pressing Shift+Enter) will list all files under the input directory
import os
for dirname, _, filenames in os.walk('/kaggle/input'):
for filename in filenames:
print(os.path.join(dirname, filename))
# You can write up to 20GB to the current directory (/kaggle/working/) that gets preserved as output when you create a version using "Save & Run All"
# You can also write temporary files to /kaggle/temp/, but they won't be saved outside of the current session
InΒ [Β ]:
InΒ [1]:
pip install ultralytics
Collecting ultralytics Downloading ultralytics-8.3.147-py3-none-any.whl.metadata (37 kB) Requirement already satisfied: numpy>=1.23.0 in /opt/conda/lib/python3.10/site-packages (from ultralytics) (1.26.4) Requirement already satisfied: matplotlib>=3.3.0 in /opt/conda/lib/python3.10/site-packages (from ultralytics) (3.7.5) Requirement already satisfied: opencv-python>=4.6.0 in /opt/conda/lib/python3.10/site-packages (from ultralytics) (4.10.0.84) Requirement already satisfied: pillow>=7.1.2 in /opt/conda/lib/python3.10/site-packages (from ultralytics) (10.3.0) Requirement already satisfied: pyyaml>=5.3.1 in /opt/conda/lib/python3.10/site-packages (from ultralytics) (6.0.2) Requirement already satisfied: requests>=2.23.0 in /opt/conda/lib/python3.10/site-packages (from ultralytics) (2.32.3) Requirement already satisfied: scipy>=1.4.1 in /opt/conda/lib/python3.10/site-packages (from ultralytics) (1.14.1) Requirement already satisfied: torch>=1.8.0 in /opt/conda/lib/python3.10/site-packages (from ultralytics) (2.4.0) Requirement already satisfied: torchvision>=0.9.0 in /opt/conda/lib/python3.10/site-packages (from ultralytics) (0.19.0) Requirement already satisfied: tqdm>=4.64.0 in /opt/conda/lib/python3.10/site-packages (from ultralytics) (4.66.4) Requirement already satisfied: psutil in /opt/conda/lib/python3.10/site-packages (from ultralytics) (5.9.3) Requirement already satisfied: py-cpuinfo in /opt/conda/lib/python3.10/site-packages (from ultralytics) (9.0.0) Requirement already satisfied: pandas>=1.1.4 in /opt/conda/lib/python3.10/site-packages (from ultralytics) (2.2.3) Collecting ultralytics-thop>=2.0.0 (from ultralytics) Downloading ultralytics_thop-2.0.14-py3-none-any.whl.metadata (9.4 kB) Requirement already satisfied: contourpy>=1.0.1 in /opt/conda/lib/python3.10/site-packages (from matplotlib>=3.3.0->ultralytics) (1.2.1) Requirement already satisfied: cycler>=0.10 in /opt/conda/lib/python3.10/site-packages (from matplotlib>=3.3.0->ultralytics) (0.12.1) Requirement already satisfied: fonttools>=4.22.0 in /opt/conda/lib/python3.10/site-packages (from matplotlib>=3.3.0->ultralytics) (4.53.0) Requirement already satisfied: kiwisolver>=1.0.1 in /opt/conda/lib/python3.10/site-packages (from matplotlib>=3.3.0->ultralytics) (1.4.5) Requirement already satisfied: packaging>=20.0 in /opt/conda/lib/python3.10/site-packages (from matplotlib>=3.3.0->ultralytics) (21.3) Requirement already satisfied: pyparsing>=2.3.1 in /opt/conda/lib/python3.10/site-packages (from matplotlib>=3.3.0->ultralytics) (3.1.2) Requirement already satisfied: python-dateutil>=2.7 in /opt/conda/lib/python3.10/site-packages (from matplotlib>=3.3.0->ultralytics) (2.9.0.post0) Requirement already satisfied: pytz>=2020.1 in /opt/conda/lib/python3.10/site-packages (from pandas>=1.1.4->ultralytics) (2024.1) Requirement already satisfied: tzdata>=2022.7 in /opt/conda/lib/python3.10/site-packages (from pandas>=1.1.4->ultralytics) (2024.1) Requirement already satisfied: charset-normalizer<4,>=2 in /opt/conda/lib/python3.10/site-packages (from requests>=2.23.0->ultralytics) (3.3.2) Requirement already satisfied: idna<4,>=2.5 in /opt/conda/lib/python3.10/site-packages (from requests>=2.23.0->ultralytics) (3.7) Requirement already satisfied: urllib3<3,>=1.21.1 in /opt/conda/lib/python3.10/site-packages (from requests>=2.23.0->ultralytics) (1.26.18) Requirement already satisfied: certifi>=2017.4.17 in /opt/conda/lib/python3.10/site-packages (from requests>=2.23.0->ultralytics) (2024.6.2) Requirement already satisfied: filelock in /opt/conda/lib/python3.10/site-packages (from torch>=1.8.0->ultralytics) (3.15.1) Requirement already satisfied: typing-extensions>=4.8.0 in /opt/conda/lib/python3.10/site-packages (from torch>=1.8.0->ultralytics) (4.12.2) Requirement already satisfied: sympy in /opt/conda/lib/python3.10/site-packages (from torch>=1.8.0->ultralytics) (1.13.3) Requirement already satisfied: networkx in /opt/conda/lib/python3.10/site-packages (from torch>=1.8.0->ultralytics) (3.3) Requirement already satisfied: jinja2 in /opt/conda/lib/python3.10/site-packages (from torch>=1.8.0->ultralytics) (3.1.4) Requirement already satisfied: fsspec in /opt/conda/lib/python3.10/site-packages (from torch>=1.8.0->ultralytics) (2024.6.0) Requirement already satisfied: six>=1.5 in /opt/conda/lib/python3.10/site-packages (from python-dateutil>=2.7->matplotlib>=3.3.0->ultralytics) (1.16.0) Requirement already satisfied: MarkupSafe>=2.0 in /opt/conda/lib/python3.10/site-packages (from jinja2->torch>=1.8.0->ultralytics) (2.1.5) Requirement already satisfied: mpmath<1.4,>=1.1.0 in /opt/conda/lib/python3.10/site-packages (from sympy->torch>=1.8.0->ultralytics) (1.3.0) Downloading ultralytics-8.3.147-py3-none-any.whl (1.0 MB) ββββββββββββββββββββββββββββββββββββββββ 1.0/1.0 MB 33.0 MB/s eta 0:00:00 Downloading ultralytics_thop-2.0.14-py3-none-any.whl (26 kB) Installing collected packages: ultralytics-thop, ultralytics Successfully installed ultralytics-8.3.147 ultralytics-thop-2.0.14 Note: you may need to restart the kernel to use updated packages.
InΒ [2]:
!pip install roboflow
from roboflow import Roboflow
rf = Roboflow(api_key="GZBlBPlRdRvB6uR7BqaS")
project = rf.workspace("detection-e83li").project("smokeandfire")
version = project.version(2)
dataset = version.download("yolov8")
Collecting roboflow Downloading roboflow-1.1.65-py3-none-any.whl.metadata (9.7 kB) Requirement already satisfied: certifi in /opt/conda/lib/python3.10/site-packages (from roboflow) (2024.6.2) Requirement already satisfied: idna==3.7 in /opt/conda/lib/python3.10/site-packages (from roboflow) (3.7) Requirement already satisfied: cycler in /opt/conda/lib/python3.10/site-packages (from roboflow) (0.12.1) Requirement already satisfied: kiwisolver>=1.3.1 in /opt/conda/lib/python3.10/site-packages (from roboflow) (1.4.5) Requirement already satisfied: matplotlib in /opt/conda/lib/python3.10/site-packages (from roboflow) (3.7.5) Requirement already satisfied: numpy>=1.18.5 in /opt/conda/lib/python3.10/site-packages (from roboflow) (1.26.4) Requirement already satisfied: opencv-python-headless==4.10.0.84 in /opt/conda/lib/python3.10/site-packages (from roboflow) (4.10.0.84) Requirement already satisfied: Pillow>=7.1.2 in /opt/conda/lib/python3.10/site-packages (from roboflow) (10.3.0) Collecting pillow-heif>=0.18.0 (from roboflow) Downloading pillow_heif-0.22.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (9.6 kB) Requirement already satisfied: python-dateutil in /opt/conda/lib/python3.10/site-packages (from roboflow) (2.9.0.post0) Requirement already satisfied: python-dotenv in /opt/conda/lib/python3.10/site-packages (from roboflow) (1.0.1) Requirement already satisfied: requests in /opt/conda/lib/python3.10/site-packages (from roboflow) (2.32.3) Requirement already satisfied: six in /opt/conda/lib/python3.10/site-packages (from roboflow) (1.16.0) Requirement already satisfied: urllib3>=1.26.6 in /opt/conda/lib/python3.10/site-packages (from roboflow) (1.26.18) Requirement already satisfied: tqdm>=4.41.0 in /opt/conda/lib/python3.10/site-packages (from roboflow) (4.66.4) Requirement already satisfied: PyYAML>=5.3.1 in /opt/conda/lib/python3.10/site-packages (from roboflow) (6.0.2) Requirement already satisfied: requests-toolbelt in /opt/conda/lib/python3.10/site-packages (from roboflow) (0.10.1) Collecting filetype (from roboflow) Downloading filetype-1.2.0-py2.py3-none-any.whl.metadata (6.5 kB) Requirement already satisfied: contourpy>=1.0.1 in /opt/conda/lib/python3.10/site-packages (from matplotlib->roboflow) (1.2.1) Requirement already satisfied: fonttools>=4.22.0 in /opt/conda/lib/python3.10/site-packages (from matplotlib->roboflow) (4.53.0) Requirement already satisfied: packaging>=20.0 in /opt/conda/lib/python3.10/site-packages (from matplotlib->roboflow) (21.3) Requirement already satisfied: pyparsing>=2.3.1 in /opt/conda/lib/python3.10/site-packages (from matplotlib->roboflow) (3.1.2) Requirement already satisfied: charset-normalizer<4,>=2 in /opt/conda/lib/python3.10/site-packages (from requests->roboflow) (3.3.2) Downloading roboflow-1.1.65-py3-none-any.whl (85 kB) ββββββββββββββββββββββββββββββββββββββββ 85.8/85.8 kB 3.9 MB/s eta 0:00:00 Downloading pillow_heif-0.22.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (7.8 MB) ββββββββββββββββββββββββββββββββββββββββ 7.8/7.8 MB 93.2 MB/s eta 0:00:00:00:0100:01 Downloading filetype-1.2.0-py2.py3-none-any.whl (19 kB) Installing collected packages: filetype, pillow-heif, roboflow Successfully installed filetype-1.2.0 pillow-heif-0.22.0 roboflow-1.1.65 loading Roboflow workspace... loading Roboflow project...
Downloading Dataset Version Zip in smokeandfire-2 to yolov8:: 100%|ββββββββββ| 381019/381019 [00:11<00:00, 32475.34it/s]
Extracting Dataset Version Zip to smokeandfire-2 in yolov8:: 100%|ββββββββββ| 19708/19708 [00:02<00:00, 8893.65it/s]
Creating new Ultralytics Settings v0.0.6 file β View Ultralytics Settings with 'yolo settings' or at '/root/.config/Ultralytics/settings.json' Update Settings with 'yolo settings key=value', i.e. 'yolo settings runs_dir=path/to/dir'. For help see https://docs.ultralytics.com/quickstart/#ultralytics-settings.
YOLO v8- MediumΒΆ
InΒ [4]:
!yolo train model=yolov8m.pt data='/kaggle/working/smokeandfire-2/data.yaml' epochs=20 imgsz=640 batch=32 augment=True
Downloading https://github.com/ultralytics/assets/releases/download/v8.3.0/yolov8m.pt to 'yolov8m.pt'... 100%|βββββββββββββββββββββββββββββββββββββββ| 49.7M/49.7M [00:00<00:00, 329MB/s] Ultralytics 8.3.70 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB) engine/trainer: task=detect, mode=train, model=yolov8m.pt, data=/kaggle/working/smokeandfire-2/data.yaml, epochs=20, time=None, patience=100, batch=32, imgsz=640, save=True, save_period=-1, cache=False, device=None, workers=8, project=None, name=train, exist_ok=False, pretrained=True, optimizer=auto, verbose=True, seed=0, deterministic=True, single_cls=False, rect=False, cos_lr=False, close_mosaic=10, resume=False, amp=True, fraction=1.0, profile=False, freeze=None, multi_scale=False, overlap_mask=True, mask_ratio=4, dropout=0.0, val=True, split=val, save_json=False, save_hybrid=False, conf=None, iou=0.7, max_det=300, half=False, dnn=False, plots=True, source=None, vid_stride=1, stream_buffer=False, visualize=False, augment=True, agnostic_nms=False, classes=None, retina_masks=False, embed=None, show=False, save_frames=False, save_txt=False, save_conf=False, save_crop=False, show_labels=True, show_conf=True, show_boxes=True, line_width=None, format=torchscript, keras=False, optimize=False, int8=False, dynamic=False, simplify=True, opset=None, workspace=None, nms=False, lr0=0.01, lrf=0.01, momentum=0.937, weight_decay=0.0005, warmup_epochs=3.0, warmup_momentum=0.8, warmup_bias_lr=0.1, box=7.5, cls=0.5, dfl=1.5, pose=12.0, kobj=1.0, nbs=64, hsv_h=0.015, hsv_s=0.7, hsv_v=0.4, degrees=0.0, translate=0.1, scale=0.5, shear=0.0, perspective=0.0, flipud=0.0, fliplr=0.5, bgr=0.0, mosaic=1.0, mixup=0.0, copy_paste=0.0, copy_paste_mode=flip, auto_augment=randaugment, erasing=0.4, crop_fraction=1.0, cfg=None, tracker=botsort.yaml, save_dir=runs/detect/train Downloading https://ultralytics.com/assets/Arial.ttf to '/root/.config/Ultralytics/Arial.ttf'... 100%|ββββββββββββββββββββββββββββββββββββββββ| 755k/755k [00:00<00:00, 22.0MB/s] Overriding model.yaml nc=80 with nc=2 from n params module arguments 0 -1 1 1392 ultralytics.nn.modules.conv.Conv [3, 48, 3, 2] 1 -1 1 41664 ultralytics.nn.modules.conv.Conv [48, 96, 3, 2] 2 -1 2 111360 ultralytics.nn.modules.block.C2f [96, 96, 2, True] 3 -1 1 166272 ultralytics.nn.modules.conv.Conv [96, 192, 3, 2] 4 -1 4 813312 ultralytics.nn.modules.block.C2f [192, 192, 4, True] 5 -1 1 664320 ultralytics.nn.modules.conv.Conv [192, 384, 3, 2] 6 -1 4 3248640 ultralytics.nn.modules.block.C2f [384, 384, 4, True] 7 -1 1 1991808 ultralytics.nn.modules.conv.Conv [384, 576, 3, 2] 8 -1 2 3985920 ultralytics.nn.modules.block.C2f [576, 576, 2, True] 9 -1 1 831168 ultralytics.nn.modules.block.SPPF [576, 576, 5] 10 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 11 [-1, 6] 1 0 ultralytics.nn.modules.conv.Concat [1] 12 -1 2 1993728 ultralytics.nn.modules.block.C2f [960, 384, 2] 13 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 14 [-1, 4] 1 0 ultralytics.nn.modules.conv.Concat [1] 15 -1 2 517632 ultralytics.nn.modules.block.C2f [576, 192, 2] 16 -1 1 332160 ultralytics.nn.modules.conv.Conv [192, 192, 3, 2] 17 [-1, 12] 1 0 ultralytics.nn.modules.conv.Concat [1] 18 -1 2 1846272 ultralytics.nn.modules.block.C2f [576, 384, 2] 19 -1 1 1327872 ultralytics.nn.modules.conv.Conv [384, 384, 3, 2] 20 [-1, 9] 1 0 ultralytics.nn.modules.conv.Concat [1] 21 -1 2 4207104 ultralytics.nn.modules.block.C2f [960, 576, 2] 22 [15, 18, 21] 1 3776854 ultralytics.nn.modules.head.Detect [2, [192, 384, 576]] Model summary: 295 layers, 25,857,478 parameters, 25,857,462 gradients, 79.1 GFLOPs Transferred 469/475 items from pretrained weights TensorBoard: Start with 'tensorboard --logdir runs/detect/train', view at http://localhost:6006/ Freezing layer 'model.22.dfl.conv.weight' AMP: running Automatic Mixed Precision (AMP) checks... Downloading https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11n.pt to 'yolo11n.pt'... 100%|ββββββββββββββββββββββββββββββββββββββ| 5.35M/5.35M [00:00<00:00, 97.3MB/s] AMP: checks passed β train: Scanning /kaggle/working/smokeandfire-2/train/labels... 8243 images, 11 b train: New cache created: /kaggle/working/smokeandfire-2/train/labels.cache WARNING β οΈ Box and segment counts should be equal, but got len(segments) = 12, len(boxes) = 16365. To resolve this only boxes will be used and all segments will be removed. To avoid this please supply either a detect or segment dataset, not a detect-segment mixed dataset. /opt/conda/lib/python3.10/site-packages/albumentations/__init__.py:24: UserWarning: A new version of Albumentations is available: 2.0.2 (you have 1.4.21). Upgrade using: pip install -U albumentations. To disable automatic update checks, set the environment variable NO_ALBUMENTATIONS_UPDATE to 1. check_for_updates() albumentations: Blur(p=0.01, blur_limit=(3, 7)), MedianBlur(p=0.01, blur_limit=(3, 7)), ToGray(p=0.01, num_output_channels=3, method='weighted_average'), CLAHE(p=0.01, clip_limit=(1.0, 4.0), tile_grid_size=(8, 8)) /opt/conda/lib/python3.10/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock. self.pid = os.fork() val: Scanning /kaggle/working/smokeandfire-2/valid/labels... 1062 images, 3 back val: New cache created: /kaggle/working/smokeandfire-2/valid/labels.cache /opt/conda/lib/python3.10/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock. self.pid = os.fork() Plotting labels to runs/detect/train/labels.jpg... optimizer: 'optimizer=auto' found, ignoring 'lr0=0.01' and 'momentum=0.937' and determining best 'optimizer', 'lr0' and 'momentum' automatically... optimizer: AdamW(lr=0.001667, momentum=0.9) with parameter groups 77 weight(decay=0.0), 84 weight(decay=0.0005), 83 bias(decay=0.0) TensorBoard: model graph visualization added β Image sizes 640 train, 640 val Using 4 dataloader workers Logging results to runs/detect/train Starting training for 20 epochs... Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 1/20 14.4G 1.478 1.91 1.611 111 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.397 0.454 0.366 0.216 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 2/20 14.9G 1.486 1.699 1.629 68 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.677 0.564 0.604 0.419 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 3/20 15G 1.425 1.561 1.586 42 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.724 0.643 0.689 0.498 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 4/20 15G 1.375 1.481 1.549 67 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.815 0.688 0.744 0.552 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 5/20 15.1G 1.312 1.359 1.501 69 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.841 0.756 0.795 0.6 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 6/20 15.1G 1.281 1.302 1.48 65 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.844 0.766 0.796 0.605 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 7/20 15.1G 1.235 1.236 1.449 66 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.858 0.798 0.813 0.628 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 8/20 15.1G 1.209 1.185 1.43 54 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.856 0.816 0.828 0.631 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 9/20 15.1G 1.175 1.136 1.412 49 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.858 0.809 0.826 0.647 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 10/20 15.1G 1.17 1.127 1.415 79 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.867 0.825 0.841 0.653 Closing dataloader mosaic albumentations: Blur(p=0.01, blur_limit=(3, 7)), MedianBlur(p=0.01, blur_limit=(3, 7)), ToGray(p=0.01, num_output_channels=3, method='weighted_average'), CLAHE(p=0.01, clip_limit=(1.0, 4.0), tile_grid_size=(8, 8)) /opt/conda/lib/python3.10/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock. self.pid = os.fork() Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 11/20 15G 1.21 1.1 1.46 41 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.873 0.835 0.837 0.652 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 12/20 15.1G 1.186 1.05 1.434 38 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.869 0.838 0.842 0.657 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 13/20 15G 1.166 1.011 1.416 34 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.879 0.829 0.841 0.664 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 14/20 15G 1.139 0.984 1.402 39 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.888 0.84 0.861 0.68 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 15/20 15.1G 1.108 0.9416 1.384 25 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.876 0.866 0.869 0.692 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 16/20 15G 1.09 0.9 1.367 35 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.897 0.853 0.863 0.69 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 17/20 15.1G 1.058 0.8724 1.347 45 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.891 0.851 0.867 0.694 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 18/20 15.1G 1.034 0.8338 1.328 33 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.887 0.86 0.872 0.703 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 19/20 15.1G 1.011 0.7975 1.306 31 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.885 0.869 0.869 0.7 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 20/20 15G 0.9823 0.7615 1.291 37 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.898 0.863 0.877 0.709 20 epochs completed in 1.651 hours. Optimizer stripped from runs/detect/train/weights/last.pt, 52.0MB Optimizer stripped from runs/detect/train/weights/best.pt, 52.0MB Validating runs/detect/train/weights/best.pt... Ultralytics 8.3.70 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB) Model summary (fused): 218 layers, 25,840,918 parameters, 0 gradients, 78.7 GFLOPs Class Images Instances Box(P R mAP50 m all 1062 2081 0.894 0.86 0.876 0.711 Fire 859 1320 0.909 0.943 0.932 0.76 Smoke 559 761 0.878 0.777 0.819 0.661 Speed: 0.1ms preprocess, 22.1ms inference, 0.0ms loss, 0.7ms postprocess per image Results saved to runs/detect/train π‘ Learn more at https://docs.ultralytics.com/modes/train
yoloV8-nanoΒΆ
InΒ [3]:
!yolo train model=yolov8n.pt data='/kaggle/working/smokeandfire-2/data.yaml' epochs=80 imgsz=640 batch=32 augment=True
Downloading https://github.com/ultralytics/assets/releases/download/v8.3.0/yolov8n.pt to 'yolov8n.pt'... 100%|βββββββββββββββββββββββββββββββββββββββ| 6.25M/6.25M [00:00<00:00, 175MB/s] Ultralytics 8.3.70 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB) engine/trainer: task=detect, mode=train, model=yolov8n.pt, data=/kaggle/working/smokeandfire-2/data.yaml, epochs=80, time=None, patience=100, batch=32, imgsz=640, save=True, save_period=-1, cache=False, device=None, workers=8, project=None, name=train, exist_ok=False, pretrained=True, optimizer=auto, verbose=True, seed=0, deterministic=True, single_cls=False, rect=False, cos_lr=False, close_mosaic=10, resume=False, amp=True, fraction=1.0, profile=False, freeze=None, multi_scale=False, overlap_mask=True, mask_ratio=4, dropout=0.0, val=True, split=val, save_json=False, save_hybrid=False, conf=None, iou=0.7, max_det=300, half=False, dnn=False, plots=True, source=None, vid_stride=1, stream_buffer=False, visualize=False, augment=True, agnostic_nms=False, classes=None, retina_masks=False, embed=None, show=False, save_frames=False, save_txt=False, save_conf=False, save_crop=False, show_labels=True, show_conf=True, show_boxes=True, line_width=None, format=torchscript, keras=False, optimize=False, int8=False, dynamic=False, simplify=True, opset=None, workspace=None, nms=False, lr0=0.01, lrf=0.01, momentum=0.937, weight_decay=0.0005, warmup_epochs=3.0, warmup_momentum=0.8, warmup_bias_lr=0.1, box=7.5, cls=0.5, dfl=1.5, pose=12.0, kobj=1.0, nbs=64, hsv_h=0.015, hsv_s=0.7, hsv_v=0.4, degrees=0.0, translate=0.1, scale=0.5, shear=0.0, perspective=0.0, flipud=0.0, fliplr=0.5, bgr=0.0, mosaic=1.0, mixup=0.0, copy_paste=0.0, copy_paste_mode=flip, auto_augment=randaugment, erasing=0.4, crop_fraction=1.0, cfg=None, tracker=botsort.yaml, save_dir=runs/detect/train Downloading https://ultralytics.com/assets/Arial.ttf to '/root/.config/Ultralytics/Arial.ttf'... 100%|ββββββββββββββββββββββββββββββββββββββββ| 755k/755k [00:00<00:00, 46.1MB/s] Overriding model.yaml nc=80 with nc=2 from n params module arguments 0 -1 1 464 ultralytics.nn.modules.conv.Conv [3, 16, 3, 2] 1 -1 1 4672 ultralytics.nn.modules.conv.Conv [16, 32, 3, 2] 2 -1 1 7360 ultralytics.nn.modules.block.C2f [32, 32, 1, True] 3 -1 1 18560 ultralytics.nn.modules.conv.Conv [32, 64, 3, 2] 4 -1 2 49664 ultralytics.nn.modules.block.C2f [64, 64, 2, True] 5 -1 1 73984 ultralytics.nn.modules.conv.Conv [64, 128, 3, 2] 6 -1 2 197632 ultralytics.nn.modules.block.C2f [128, 128, 2, True] 7 -1 1 295424 ultralytics.nn.modules.conv.Conv [128, 256, 3, 2] 8 -1 1 460288 ultralytics.nn.modules.block.C2f [256, 256, 1, True] 9 -1 1 164608 ultralytics.nn.modules.block.SPPF [256, 256, 5] 10 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 11 [-1, 6] 1 0 ultralytics.nn.modules.conv.Concat [1] 12 -1 1 148224 ultralytics.nn.modules.block.C2f [384, 128, 1] 13 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 14 [-1, 4] 1 0 ultralytics.nn.modules.conv.Concat [1] 15 -1 1 37248 ultralytics.nn.modules.block.C2f [192, 64, 1] 16 -1 1 36992 ultralytics.nn.modules.conv.Conv [64, 64, 3, 2] 17 [-1, 12] 1 0 ultralytics.nn.modules.conv.Concat [1] 18 -1 1 123648 ultralytics.nn.modules.block.C2f [192, 128, 1] 19 -1 1 147712 ultralytics.nn.modules.conv.Conv [128, 128, 3, 2] 20 [-1, 9] 1 0 ultralytics.nn.modules.conv.Concat [1] 21 -1 1 493056 ultralytics.nn.modules.block.C2f [384, 256, 1] 22 [15, 18, 21] 1 751702 ultralytics.nn.modules.head.Detect [2, [64, 128, 256]] Model summary: 225 layers, 3,011,238 parameters, 3,011,222 gradients, 8.2 GFLOPs Transferred 319/355 items from pretrained weights TensorBoard: Start with 'tensorboard --logdir runs/detect/train', view at http://localhost:6006/ Freezing layer 'model.22.dfl.conv.weight' AMP: running Automatic Mixed Precision (AMP) checks... Downloading https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11n.pt to 'yolo11n.pt'... 100%|βββββββββββββββββββββββββββββββββββββββ| 5.35M/5.35M [00:00<00:00, 171MB/s] AMP: checks passed β train: Scanning /kaggle/working/smokeandfire-2/train/labels... 8243 images, 11 b train: New cache created: /kaggle/working/smokeandfire-2/train/labels.cache WARNING β οΈ Box and segment counts should be equal, but got len(segments) = 12, len(boxes) = 16365. To resolve this only boxes will be used and all segments will be removed. To avoid this please supply either a detect or segment dataset, not a detect-segment mixed dataset. /opt/conda/lib/python3.10/site-packages/albumentations/__init__.py:24: UserWarning: A new version of Albumentations is available: 2.0.3 (you have 1.4.21). Upgrade using: pip install -U albumentations. To disable automatic update checks, set the environment variable NO_ALBUMENTATIONS_UPDATE to 1. check_for_updates() albumentations: Blur(p=0.01, blur_limit=(3, 7)), MedianBlur(p=0.01, blur_limit=(3, 7)), ToGray(p=0.01, num_output_channels=3, method='weighted_average'), CLAHE(p=0.01, clip_limit=(1.0, 4.0), tile_grid_size=(8, 8)) /opt/conda/lib/python3.10/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock. self.pid = os.fork() val: Scanning /kaggle/working/smokeandfire-2/valid/labels... 1062 images, 3 back val: New cache created: /kaggle/working/smokeandfire-2/valid/labels.cache /opt/conda/lib/python3.10/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock. self.pid = os.fork() Plotting labels to runs/detect/train/labels.jpg... optimizer: 'optimizer=auto' found, ignoring 'lr0=0.01' and 'momentum=0.937' and determining best 'optimizer', 'lr0' and 'momentum' automatically... optimizer: SGD(lr=0.01, momentum=0.9) with parameter groups 57 weight(decay=0.0), 64 weight(decay=0.0005), 63 bias(decay=0.0) TensorBoard: model graph visualization added β Image sizes 640 train, 640 val Using 4 dataloader workers Logging results to runs/detect/train Starting training for 80 epochs... Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 1/80 4.28G 1.46 2.732 1.525 111 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.643 0.594 0.623 0.448 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 2/80 4.04G 1.351 1.809 1.427 68 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.786 0.643 0.693 0.481 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 3/80 4.04G 1.395 1.662 1.462 42 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.68 0.645 0.65 0.438 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 4/80 4.04G 1.443 1.629 1.516 67 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.773 0.648 0.703 0.507 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 5/80 4.04G 1.391 1.488 1.474 69 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.832 0.706 0.757 0.562 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 6/80 4.04G 1.37 1.43 1.456 65 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.824 0.759 0.772 0.556 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 7/80 4.04G 1.32 1.352 1.423 66 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.837 0.76 0.795 0.58 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 8/80 4.04G 1.305 1.314 1.413 54 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.848 0.791 0.806 0.598 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 9/80 4.13G 1.273 1.264 1.398 49 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.834 0.815 0.82 0.616 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 10/80 4.04G 1.276 1.265 1.403 79 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.847 0.778 0.811 0.614 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 11/80 4.04G 1.234 1.212 1.376 59 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.843 0.79 0.811 0.619 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 12/80 4.04G 1.231 1.203 1.376 81 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.859 0.82 0.835 0.642 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 13/80 4.04G 1.228 1.182 1.371 57 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.852 0.796 0.815 0.633 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 14/80 4.03G 1.209 1.163 1.361 66 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.847 0.83 0.84 0.643 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 15/80 4.03G 1.193 1.14 1.357 78 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.856 0.817 0.82 0.632 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 16/80 4.04G 1.178 1.11 1.345 68 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.868 0.821 0.841 0.648 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 17/80 4.04G 1.186 1.121 1.344 68 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.862 0.832 0.847 0.659 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 18/80 4.04G 1.157 1.088 1.323 75 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.858 0.814 0.837 0.654 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 19/80 4.04G 1.166 1.096 1.331 53 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.888 0.822 0.851 0.666 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 20/80 4.04G 1.149 1.065 1.325 82 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.858 0.833 0.848 0.668 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 21/80 4.04G 1.148 1.064 1.324 62 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.863 0.828 0.841 0.664 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 22/80 4.04G 1.147 1.049 1.315 79 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.877 0.836 0.854 0.671 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 23/80 4.04G 1.124 1.029 1.307 71 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.887 0.821 0.848 0.67 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 24/80 4.04G 1.125 1.027 1.299 75 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.874 0.831 0.853 0.668 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 25/80 4.03G 1.112 1.01 1.29 51 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.872 0.839 0.848 0.673 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 26/80 4.04G 1.116 1.014 1.295 81 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.869 0.836 0.852 0.677 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 27/80 4.03G 1.105 0.9947 1.291 73 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.891 0.836 0.851 0.679 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 28/80 4.03G 1.098 0.9788 1.283 58 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.839 0.868 0.684 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 29/80 4.03G 1.099 0.9862 1.29 68 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.883 0.843 0.857 0.681 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 30/80 4.04G 1.097 0.9711 1.28 40 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.889 0.837 0.861 0.686 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 31/80 4.03G 1.088 0.9652 1.276 47 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.878 0.844 0.862 0.691 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 32/80 4.03G 1.072 0.9448 1.266 68 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.886 0.833 0.857 0.684 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 33/80 4.03G 1.078 0.9493 1.263 77 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.884 0.846 0.863 0.686 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 34/80 4.04G 1.059 0.9343 1.257 47 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.892 0.843 0.86 0.688 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 35/80 4.03G 1.064 0.9361 1.264 54 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.891 0.844 0.862 0.688 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 36/80 4.04G 1.073 0.9309 1.262 70 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.888 0.844 0.862 0.686 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 37/80 4.04G 1.062 0.9132 1.257 41 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.887 0.854 0.868 0.694 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 38/80 4.03G 1.041 0.9034 1.244 62 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.891 0.846 0.861 0.693 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 39/80 4.04G 1.042 0.8871 1.243 56 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.897 0.854 0.866 0.696 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 40/80 4.03G 1.035 0.8803 1.239 77 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.842 0.863 0.693 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 41/80 4.04G 1.036 0.8864 1.238 54 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.888 0.856 0.864 0.694 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 42/80 4.04G 1.031 0.8757 1.235 61 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.894 0.855 0.867 0.696 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 43/80 4.03G 1.032 0.8717 1.238 56 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.884 0.862 0.87 0.704 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 44/80 4.03G 1.025 0.8757 1.234 65 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.89 0.86 0.866 0.701 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 45/80 4.03G 1.018 0.8574 1.232 61 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.893 0.855 0.865 0.698 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 46/80 4.04G 1.014 0.8478 1.222 74 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.898 0.849 0.863 0.7 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 47/80 4.04G 1.001 0.8449 1.217 55 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.894 0.853 0.862 0.701 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 48/80 4.03G 1.022 0.8509 1.23 63 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.893 0.853 0.868 0.703 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 49/80 4.03G 1.001 0.8292 1.214 74 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.894 0.85 0.866 0.701 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 50/80 4.03G 0.9953 0.8194 1.213 70 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.901 0.853 0.864 0.702 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 51/80 4.03G 0.9899 0.8199 1.204 69 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.892 0.856 0.866 0.704 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 52/80 4.03G 0.9847 0.811 1.204 71 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.894 0.857 0.869 0.707 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 53/80 4.03G 0.9856 0.8148 1.204 56 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.899 0.854 0.867 0.707 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 54/80 4.03G 0.9765 0.7975 1.198 79 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.854 0.871 0.709 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 55/80 4.03G 0.972 0.7906 1.196 59 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.904 0.855 0.872 0.707 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 56/80 4.03G 0.9653 0.7835 1.193 76 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.899 0.857 0.87 0.709 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 57/80 4.03G 0.9713 0.7852 1.191 63 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.903 0.854 0.87 0.709 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 58/80 4.03G 0.9614 0.7743 1.188 98 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.899 0.863 0.873 0.713 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 59/80 4.03G 0.9609 0.7649 1.19 120 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.904 0.855 0.874 0.709 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 60/80 4.03G 0.9524 0.7774 1.187 73 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.905 0.857 0.869 0.711 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 61/80 4.03G 0.9534 0.7619 1.185 63 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.898 0.856 0.868 0.707 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 62/80 4.03G 0.9391 0.7483 1.172 63 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.904 0.85 0.868 0.707 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 63/80 4.04G 0.9416 0.7508 1.178 73 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.899 0.858 0.869 0.708 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 64/80 4.04G 0.9248 0.735 1.167 62 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.9 0.861 0.872 0.71 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 65/80 4.03G 0.9385 0.7398 1.172 97 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.901 0.86 0.869 0.71 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 66/80 4.03G 0.9379 0.737 1.173 56 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.9 0.858 0.87 0.708 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 67/80 4.03G 0.9173 0.7124 1.16 79 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.906 0.859 0.871 0.71 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 68/80 4.03G 0.9239 0.7183 1.165 78 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.906 0.864 0.872 0.712 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 69/80 4.03G 0.9043 0.7065 1.153 68 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.904 0.856 0.873 0.713 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 70/80 4.03G 0.908 0.7109 1.155 75 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.902 0.857 0.871 0.71 Closing dataloader mosaic albumentations: Blur(p=0.01, blur_limit=(3, 7)), MedianBlur(p=0.01, blur_limit=(3, 7)), ToGray(p=0.01, num_output_channels=3, method='weighted_average'), CLAHE(p=0.01, clip_limit=(1.0, 4.0), tile_grid_size=(8, 8)) /opt/conda/lib/python3.10/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock. self.pid = os.fork() Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 71/80 4.27G 0.9224 0.6481 1.167 44 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.908 0.852 0.874 0.712 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 72/80 4.03G 0.9097 0.6335 1.159 34 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.901 0.861 0.873 0.714 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 73/80 4.03G 0.8956 0.6201 1.145 37 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.908 0.854 0.872 0.714 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 74/80 4.03G 0.8915 0.6142 1.146 39 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.902 0.855 0.871 0.714 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 75/80 4.03G 0.8798 0.6015 1.14 36 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.9 0.858 0.871 0.714 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 76/80 4.04G 0.8763 0.5909 1.135 32 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.9 0.857 0.871 0.715 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 77/80 4.04G 0.8724 0.5878 1.134 49 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.9 0.856 0.87 0.714 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 78/80 4.03G 0.8691 0.5789 1.13 43 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.902 0.856 0.87 0.713 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 79/80 4.03G 0.8659 0.5741 1.128 39 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.897 0.86 0.87 0.714 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 80/80 4.03G 0.852 0.5651 1.121 49 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.9 0.857 0.87 0.714 80 epochs completed in 1.898 hours. Optimizer stripped from runs/detect/train/weights/last.pt, 6.3MB Optimizer stripped from runs/detect/train/weights/best.pt, 6.3MB Validating runs/detect/train/weights/best.pt... Ultralytics 8.3.70 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB) Model summary (fused): 168 layers, 3,006,038 parameters, 0 gradients, 8.1 GFLOPs Class Images Instances Box(P R mAP50 m all 1062 2081 0.896 0.862 0.87 0.713 Fire 859 1320 0.916 0.94 0.932 0.773 Smoke 559 761 0.876 0.783 0.808 0.652 Speed: 0.1ms preprocess, 4.8ms inference, 0.0ms loss, 0.8ms postprocess per image Results saved to runs/detect/train π‘ Learn more at https://docs.ultralytics.com/modes/train
yolo V8-SmallΒΆ
InΒ [Β ]:
!yolo train model=yolov8s.pt data='/kaggle/working/smokeandfire-2/data.yaml' epochs=80 imgsz=640 batch=32 augment=True
Downloading https://github.com/ultralytics/assets/releases/download/v8.3.0/yolov8s.pt to 'yolov8s.pt'... 100%|βββββββββββββββββββββββββββββββββββββββ| 21.5M/21.5M [00:00<00:00, 186MB/s] Ultralytics 8.3.27 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB) engine/trainer: task=detect, mode=train, model=yolov8s.pt, data=/kaggle/working/smokeandfire-2/data.yaml, epochs=80, time=None, patience=100, batch=32, imgsz=640, save=True, save_period=-1, cache=False, device=None, workers=8, project=None, name=train, exist_ok=False, pretrained=True, optimizer=auto, verbose=True, seed=0, deterministic=True, single_cls=False, rect=False, cos_lr=False, close_mosaic=10, resume=False, amp=True, fraction=1.0, profile=False, freeze=None, multi_scale=False, overlap_mask=True, mask_ratio=4, dropout=0.0, val=True, split=val, save_json=False, save_hybrid=False, conf=None, iou=0.7, max_det=300, half=False, dnn=False, plots=True, source=None, vid_stride=1, stream_buffer=False, visualize=False, augment=True, agnostic_nms=False, classes=None, retina_masks=False, embed=None, show=False, save_frames=False, save_txt=False, save_conf=False, save_crop=False, show_labels=True, show_conf=True, show_boxes=True, line_width=None, format=torchscript, keras=False, optimize=False, int8=False, dynamic=False, simplify=True, opset=None, workspace=4, nms=False, lr0=0.01, lrf=0.01, momentum=0.937, weight_decay=0.0005, warmup_epochs=3.0, warmup_momentum=0.8, warmup_bias_lr=0.1, box=7.5, cls=0.5, dfl=1.5, pose=12.0, kobj=1.0, label_smoothing=0.0, nbs=64, hsv_h=0.015, hsv_s=0.7, hsv_v=0.4, degrees=0.0, translate=0.1, scale=0.5, shear=0.0, perspective=0.0, flipud=0.0, fliplr=0.5, bgr=0.0, mosaic=1.0, mixup=0.0, copy_paste=0.0, copy_paste_mode=flip, auto_augment=randaugment, erasing=0.4, crop_fraction=1.0, cfg=None, tracker=botsort.yaml, save_dir=runs/detect/train Downloading https://ultralytics.com/assets/Arial.ttf to '/root/.config/Ultralytics/Arial.ttf'... 100%|ββββββββββββββββββββββββββββββββββββββββ| 755k/755k [00:00<00:00, 22.1MB/s] Overriding model.yaml nc=80 with nc=2 from n params module arguments 0 -1 1 928 ultralytics.nn.modules.conv.Conv [3, 32, 3, 2] 1 -1 1 18560 ultralytics.nn.modules.conv.Conv [32, 64, 3, 2] 2 -1 1 29056 ultralytics.nn.modules.block.C2f [64, 64, 1, True] 3 -1 1 73984 ultralytics.nn.modules.conv.Conv [64, 128, 3, 2] 4 -1 2 197632 ultralytics.nn.modules.block.C2f [128, 128, 2, True] 5 -1 1 295424 ultralytics.nn.modules.conv.Conv [128, 256, 3, 2] 6 -1 2 788480 ultralytics.nn.modules.block.C2f [256, 256, 2, True] 7 -1 1 1180672 ultralytics.nn.modules.conv.Conv [256, 512, 3, 2] 8 -1 1 1838080 ultralytics.nn.modules.block.C2f [512, 512, 1, True] 9 -1 1 656896 ultralytics.nn.modules.block.SPPF [512, 512, 5] 10 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 11 [-1, 6] 1 0 ultralytics.nn.modules.conv.Concat [1] 12 -1 1 591360 ultralytics.nn.modules.block.C2f [768, 256, 1] 13 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 14 [-1, 4] 1 0 ultralytics.nn.modules.conv.Concat [1] 15 -1 1 148224 ultralytics.nn.modules.block.C2f [384, 128, 1] 16 -1 1 147712 ultralytics.nn.modules.conv.Conv [128, 128, 3, 2] 17 [-1, 12] 1 0 ultralytics.nn.modules.conv.Concat [1] 18 -1 1 493056 ultralytics.nn.modules.block.C2f [384, 256, 1] 19 -1 1 590336 ultralytics.nn.modules.conv.Conv [256, 256, 3, 2] 20 [-1, 9] 1 0 ultralytics.nn.modules.conv.Concat [1] 21 -1 1 1969152 ultralytics.nn.modules.block.C2f [768, 512, 1] 22 [15, 18, 21] 1 2116822 ultralytics.nn.modules.head.Detect [2, [128, 256, 512]] Model summary: 225 layers, 11,136,374 parameters, 11,136,358 gradients, 28.6 GFLOPs Transferred 349/355 items from pretrained weights TensorBoard: Start with 'tensorboard --logdir runs/detect/train', view at http://localhost:6006/ Freezing layer 'model.22.dfl.conv.weight' AMP: running Automatic Mixed Precision (AMP) checks... Downloading https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11n.pt to 'yolo11n.pt'... 100%|ββββββββββββββββββββββββββββββββββββββ| 5.35M/5.35M [00:00<00:00, 91.4MB/s] AMP: checks passed β train: Scanning /kaggle/working/smokeandfire-2/train/labels... 8243 images, 11 b train: New cache created: /kaggle/working/smokeandfire-2/train/labels.cache WARNING β οΈ Box and segment counts should be equal, but got len(segments) = 12, len(boxes) = 16365. To resolve this only boxes will be used and all segments will be removed. To avoid this please supply either a detect or segment dataset, not a detect-segment mixed dataset. /opt/conda/lib/python3.10/site-packages/albumentations/__init__.py:13: UserWarning: A new version of Albumentations is available: 1.4.21 (you have 1.4.17). Upgrade using: pip install -U albumentations. To disable automatic update checks, set the environment variable NO_ALBUMENTATIONS_UPDATE to 1. check_for_updates() albumentations: Blur(p=0.01, blur_limit=(3, 7)), MedianBlur(p=0.01, blur_limit=(3, 7)), ToGray(p=0.01, num_output_channels=3, method='weighted_average'), CLAHE(p=0.01, clip_limit=(1, 4.0), tile_grid_size=(8, 8)) /opt/conda/lib/python3.10/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock. self.pid = os.fork() val: Scanning /kaggle/working/smokeandfire-2/valid/labels... 1062 images, 3 back val: New cache created: /kaggle/working/smokeandfire-2/valid/labels.cache Plotting labels to runs/detect/train/labels.jpg... optimizer: 'optimizer=auto' found, ignoring 'lr0=0.01' and 'momentum=0.937' and determining best 'optimizer', 'lr0' and 'momentum' automatically... optimizer: SGD(lr=0.01, momentum=0.9) with parameter groups 57 weight(decay=0.0), 64 weight(decay=0.0005), 63 bias(decay=0.0) TensorBoard: model graph visualization added β Image sizes 640 train, 640 val Using 4 dataloader workers Logging results to runs/detect/train Starting training for 80 epochs... Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 1/80 8.06G 1.457 2.109 1.552 64 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.835 0.704 0.742 0.55 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 2/80 8.22G 1.255 1.287 1.389 67 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.804 0.746 0.779 0.581 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 3/80 8.28G 1.324 1.345 1.428 66 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.661 0.556 0.565 0.398 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 4/80 8.31G 1.384 1.435 1.471 70 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.735 0.629 0.682 0.502 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 5/80 8.31G 1.351 1.367 1.447 72 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.821 0.681 0.726 0.543 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 6/80 8.3G 1.304 1.301 1.434 65 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.829 0.757 0.785 0.59 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 7/80 8.29G 1.267 1.234 1.409 58 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.816 0.791 0.816 0.62 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 8/80 8.38G 1.252 1.204 1.403 77 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.832 0.754 0.778 0.599 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 9/80 8.37G 1.225 1.159 1.379 86 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.863 0.807 0.829 0.635 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 10/80 8.37G 1.229 1.166 1.381 50 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.834 0.776 0.804 0.62 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 11/80 8.29G 1.197 1.12 1.363 54 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.856 0.822 0.837 0.643 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 12/80 8.37G 1.179 1.097 1.352 64 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.83 0.806 0.814 0.63 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 13/80 8.37G 1.177 1.084 1.352 82 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.86 0.825 0.83 0.648 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 14/80 8.38G 1.159 1.055 1.335 63 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.872 0.823 0.832 0.65 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 15/80 8.37G 1.148 1.04 1.329 64 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.871 0.828 0.838 0.656 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 16/80 8.37G 1.123 0.9982 1.309 70 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.88 0.838 0.851 0.668 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 17/80 8.38G 1.117 1.002 1.317 62 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.882 0.814 0.853 0.669 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 18/80 8.29G 1.106 0.9717 1.296 57 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.877 0.84 0.841 0.664 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 19/80 8.29G 1.106 0.9747 1.3 46 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.873 0.838 0.85 0.676 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 20/80 8.29G 1.085 0.9517 1.287 57 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.863 0.838 0.859 0.676 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 21/80 8.37G 1.086 0.9434 1.281 56 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.879 0.84 0.856 0.679 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 22/80 8.37G 1.077 0.9353 1.278 48 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.884 0.832 0.86 0.681 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 23/80 8.37G 1.078 0.9254 1.273 57 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.898 0.843 0.863 0.687 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 24/80 8.37G 1.069 0.9217 1.273 79 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.891 0.843 0.858 0.685 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 25/80 8.37G 1.062 0.9139 1.268 63 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.885 0.844 0.867 0.691 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 26/80 8.37G 1.05 0.8975 1.258 71 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.881 0.844 0.864 0.688 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 27/80 8.3G 1.034 0.88 1.25 104 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.887 0.857 0.864 0.688 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 28/80 8.3G 1.038 0.8747 1.255 46 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.882 0.858 0.874 0.701 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 29/80 8.37G 1.023 0.8579 1.246 67 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.87 0.861 0.864 0.696 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 30/80 8.37G 1.022 0.8496 1.242 73 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.888 0.849 0.866 0.694 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 31/80 8.37G 1.015 0.8321 1.239 64 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.887 0.849 0.86 0.699 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 32/80 8.37G 0.9935 0.8153 1.218 80 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.892 0.851 0.869 0.704 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 33/80 8.37G 1.003 0.8232 1.226 74 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.906 0.848 0.87 0.706 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 34/80 8.37G 0.996 0.8132 1.228 55 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.888 0.855 0.866 0.7 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 35/80 8.37G 0.9954 0.8055 1.225 59 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.888 0.857 0.867 0.7 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 36/80 8.37G 0.987 0.7936 1.213 60 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.893 0.849 0.867 0.703 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 37/80 8.37G 0.9708 0.78 1.205 58 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.897 0.858 0.873 0.709 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 38/80 8.3G 0.9749 0.7767 1.199 71 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.898 0.86 0.873 0.706 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 39/80 8.28G 0.9713 0.7719 1.198 55 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.892 0.857 0.869 0.704 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 40/80 8.3G 0.9569 0.7549 1.193 63 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.901 0.858 0.87 0.705 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 41/80 8.29G 0.9433 0.7394 1.185 89 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.902 0.859 0.872 0.711 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 43/80 8.37G 0.9451 0.7332 1.183 41 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.888 0.86 0.871 0.709 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 46/80 8.37G 0.9217 0.7077 1.166 82 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.897 0.86 0.875 0.713 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 47/80 8.3G 0.9227 0.6978 1.159 60 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.892 0.857 0.879 0.714 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 48/80 8.29G 0.9125 0.6868 1.166 59 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.859 0.872 0.709 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 49/80 8.29G 0.9072 0.6779 1.153 41 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.896 0.862 0.877 0.716 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 50/80 8.28G 0.9092 0.6773 1.159 48 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.896 0.862 0.874 0.715 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 51/80 8.38G 0.8942 0.6712 1.147 74 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.901 0.857 0.873 0.711 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 52/80 8.29G 0.8948 0.668 1.144 63 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.901 0.858 0.876 0.714 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 53/80 8.29G 0.8872 0.6579 1.144 67 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.894 0.866 0.875 0.712 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 54/80 8.29G 0.8826 0.6462 1.141 68 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.898 0.866 0.878 0.717 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 55/80 8.29G 0.8757 0.6416 1.134 53 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.898 0.86 0.876 0.717 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 56/80 8.3G 0.8757 0.6442 1.137 61 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.889 0.854 0.876 0.718 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 57/80 8.28G 0.8671 0.6283 1.126 87 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.898 0.856 0.875 0.725 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 58/80 8.28G 0.8548 0.6194 1.12 52 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.896 0.858 0.877 0.721 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 59/80 8.37G 0.8509 0.6142 1.115 49 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.9 0.855 0.874 0.716 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 60/80 8.29G 0.8487 0.6082 1.119 71 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.89 0.853 0.873 0.718 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 61/80 8.29G 0.8438 0.5988 1.115 83 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.901 0.85 0.871 0.715 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 62/80 8.29G 0.8404 0.5922 1.116 60 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.896 0.859 0.873 0.718 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 63/80 8.29G 0.827 0.5874 1.109 59 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.861 0.873 0.719 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 64/80 8.37G 0.8218 0.5798 1.103 55 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.89 0.867 0.871 0.719 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 65/80 8.37G 0.83 0.5832 1.105 79 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.893 0.862 0.874 0.719 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 66/80 8.37G 0.8059 0.5597 1.089 74 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.896 0.863 0.875 0.721 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 67/80 8.37G 0.8043 0.5658 1.092 96 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.893 0.865 0.872 0.72 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 68/80 8.37G 0.7967 0.5554 1.09 85 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.886 0.866 0.874 0.721 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 69/80 8.37G 0.7953 0.5495 1.088 71 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.89 0.861 0.874 0.722 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 70/80 8.37G 0.796 0.5458 1.081 56 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.891 0.863 0.87 0.718 Closing dataloader mosaic albumentations: Blur(p=0.01, blur_limit=(3, 7)), MedianBlur(p=0.01, blur_limit=(3, 7)), ToGray(p=0.01, num_output_channels=3, method='weighted_average'), CLAHE(p=0.01, clip_limit=(1, 4.0), tile_grid_size=(8, 8)) /opt/conda/lib/python3.10/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock. self.pid = os.fork() Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 71/80 8.37G 0.7757 0.4712 1.066 46 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.896 0.856 0.867 0.717 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 72/80 8.37G 0.7661 0.4539 1.054 34 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.898 0.857 0.868 0.715 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 73/80 8.37G 0.7474 0.4406 1.044 35 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.894 0.856 0.868 0.718 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 74/80 8.37G 0.7435 0.4356 1.043 39 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.89 0.86 0.871 0.72 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 75/80 8.37G 0.7309 0.4248 1.037 37 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.886 0.862 0.871 0.719 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 76/80 8.37G 0.7278 0.4191 1.033 32 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.888 0.86 0.871 0.719 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 77/80 8.37G 0.7142 0.4116 1.026 49 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.892 0.859 0.872 0.72 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 78/80 8.37G 0.7086 0.4014 1.025 42 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.893 0.857 0.87 0.721 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 79/80 8.37G 0.701 0.3992 1.019 37 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.891 0.854 0.869 0.721 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 80/80 8.37G 0.6978 0.3968 1.019 46 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.889 0.854 0.869 0.721 80 epochs completed in 3.195 hours. Optimizer stripped from runs/detect/train/weights/last.pt, 22.5MB Optimizer stripped from runs/detect/train/weights/best.pt, 22.5MB Validating runs/detect/train/weights/best.pt... Ultralytics 8.3.27 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB) Model summary (fused): 168 layers, 11,126,358 parameters, 0 gradients, 28.4 GFLOPs Class Images Instances Box(P R mAP50 m all 1062 2081 0.878 0.867 0.874 0.718 Fire 859 1320 0.903 0.943 0.936 0.783 Smoke 559 761 0.853 0.792 0.812 0.653 Speed: 0.1ms preprocess, 10.0ms inference, 0.0ms loss, 0.7ms postprocess per image Results saved to runs/detect/train π‘ Learn more at https://docs.ultralytics.com/modes/train
using yolo8s cuz of unstability of yolo8m
results of yolov8sΒΆ
InΒ [Β ]:
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
img = mpimg.imread('/kaggle/working/runs/detect/train/results.png')
imgplot = plt.imshow(img)
plt.show()
InΒ [Β ]:
img = mpimg.imread('/kaggle/working/runs/detect/train/confusion_matrix_normalized.png')
imgplot = plt.imshow(img)
plt.show()
InΒ [Β ]:
img = mpimg.imread('/kaggle/working/runs/detect/train/P_curve.png')
imgplot = plt.imshow(img)
plt.show()
InΒ [Β ]:
img = mpimg.imread('/kaggle/working/runs/detect/train/val_batch0_pred.jpg')
imgplot = plt.imshow(img)
plt.show()
InΒ [Β ]:
!yolo task=detect mode=predict model="/kaggle/working/runs/detect/train/weights/best.pt" device=0 source="/kaggle/working/smokeandfire-2/test/images" save=True
InΒ [Β ]:
from glob import glob
from PIL import Image
import numpy as np
img_path = glob('/kaggle/working/runs/detect/predict/*.jpg')
img_path = np.random.choice(img_path, 10)
for image_path in img_path:
plt.imshow(Image.open(image_path))
plt.axis("off")
plt.show()
print("\n")
InΒ [Β ]:
!yolo predict model="/kaggle/working/runs/detect/train/weights/best.pt", source="/kaggle/input/fireee/fire.109.jpg"
WARNING β οΈ argument 'model=/kaggle/working/runs/detect/train/weights/best.pt,' does not require trailing comma ',', updating to 'model=/kaggle/working/runs/detect/train/weights/best.pt'.
Ultralytics 8.3.27 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB)
Model summary (fused): 168 layers, 11,126,358 parameters, 0 gradients, 28.4 GFLOPs
image 1/1 /kaggle/input/fireee/fire.109.jpg: 384x640 3 Fires, 54.8ms
Speed: 4.4ms preprocess, 54.8ms inference, 165.3ms postprocess per image at shape (1, 3, 384, 640)
Results saved to runs/detect/predict2
π‘ Learn more at https://docs.ultralytics.com/modes/predict
test on new imagesΒΆ
InΒ [Β ]:
import cv2
import matplotlib.pyplot as plt
# Path to the output image
output_image_path = "/kaggle/working/runs/detect/predict2/fire.109.jpg"
# Load the output image
output_image = cv2.imread(output_image_path)
output_image = cv2.cvtColor(output_image, cv2.COLOR_BGR2RGB)
# Display the output image
plt.figure(figsize=(8, 8))
plt.imshow(output_image)
plt.title('Detected Image - fire.109.jpg')
plt.axis('off') # Hide axes
plt.show()
InΒ [Β ]:
!yolo predict model="/kaggle/working/runs/detect/train/weights/best.pt", source="/kaggle/input/smokee/17_jpg.rf.001aef48a2d04d48f389f096c8f3e478.jpg"
InΒ [Β ]:
output_image_path = "/kaggle/working/runs/detect/predict3/17_jpg.rf.001aef48a2d04d48f389f096c8f3e478.jpg"
# Load the output image
output_image = cv2.imread(output_image_path)
output_image = cv2.cvtColor(output_image, cv2.COLOR_BGR2RGB)
# Display the output image
plt.figure(figsize=(8, 8))
plt.imshow(output_image)
plt.title('Detected Image - fire.109.jpg')
plt.axis('off') # Hide axes
plt.show()
InΒ [Β ]:
!yolo predict model="/kaggle/working/runs/detect/train/weights/best.pt", source="/kaggle/input/fire-and-smoke-detection-for-yolov8/train/images/0058_jpg.rf.629dad210eae7f57976d164974befbe2.jpg"
WARNING β οΈ argument 'model=/kaggle/working/runs/detect/train/weights/best.pt,' does not require trailing comma ',', updating to 'model=/kaggle/working/runs/detect/train/weights/best.pt'.
Ultralytics 8.3.27 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB)
Model summary (fused): 168 layers, 11,126,358 parameters, 0 gradients, 28.4 GFLOPs
image 1/1 /kaggle/input/fire-and-smoke-detection-for-yolov8/train/images/0058_jpg.rf.629dad210eae7f57976d164974befbe2.jpg: 640x640 1 Fire, 1 Smoke, 8.5ms
Speed: 4.9ms preprocess, 8.5ms inference, 159.6ms postprocess per image at shape (1, 3, 640, 640)
Results saved to runs/detect/predict4
π‘ Learn more at https://docs.ultralytics.com/modes/predict
InΒ [Β ]:
output_image_path = "/kaggle/working/runs/detect/predict4/0058_jpg.rf.629dad210eae7f57976d164974befbe2.jpg"
# Load the output image
output_image = cv2.imread(output_image_path)
output_image = cv2.cvtColor(output_image, cv2.COLOR_BGR2RGB)
# Display the output image
plt.figure(figsize=(8, 8))
plt.imshow(output_image)
plt.title('Detected Image')
plt.axis('off') # Hide axes
plt.show()
yoloV5ΒΆ
InΒ [Β ]:
!pip install roboflow
from roboflow import Roboflow
rf = Roboflow(api_key="GZBlBPlRdRvB6uR7BqaS")
project = rf.workspace("detection-e83li").project("smokeandfire")
version = project.version(2)
dataset = version.download("yolov5")
InΒ [Β ]:
!git clone https://github.com/ultralytics/yolov5.git
Cloning into 'yolov5'... remote: Enumerating objects: 17022, done. remote: Total 17022 (delta 0), reused 0 (delta 0), pack-reused 17022 (from 1) Receiving objects: 100% (17022/17022), 15.62 MiB | 25.10 MiB/s, done. Resolving deltas: 100% (11695/11695), done.
InΒ [Β ]:
%cd yolov5
/kaggle/working/yolov5
InΒ [Β ]:
!pip install -r requirements.txt
InΒ [Β ]:
%env WANDB_MODE=disabled
env: WANDB_MODE=disabled
InΒ [Β ]:
!python train.py --weights yolov5n.pt --data '/kaggle/working/yolov5/smokeandfire-2/data.yaml' --epochs 80 --imgsz 640 --batch-size 32
InΒ [Β ]:
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
img = mpimg.imread('/kaggle/working/yolov5/runs/train/exp/confusion_matrix.png')
imgplot = plt.imshow(img)
plt.show()
InΒ [Β ]:
img = mpimg.imread('/kaggle/working/yolov5/runs/train/exp/P_curve.png')
imgplot = plt.imshow(img)
plt.show()
InΒ [Β ]:
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
img = mpimg.imread('/kaggle/working/yolov5/runs/train/exp/results.png')
imgplot = plt.imshow(img)
plt.show()
InΒ [Β ]:
!python detect.py --weights /kaggle/working/yolov5/runs/train/exp/weights/best.pt --img 640 --conf 0.35 --source /kaggle/working/yolov5/smokeandfire-2/test/images
InΒ [Β ]:
# Running YOLOv5 prediction from the terminal
!python detect.py --weights /kaggle/working/yolov5/runs/train/exp/weights/best.pt --source /kaggle/input/fireee/fire.109.jpg --img 640 --conf 0.25
detect: weights=['/kaggle/working/yolov5/runs/train/exp/weights/best.pt'], source=/kaggle/input/fireee/fire.109.jpg, data=data/coco128.yaml, imgsz=[640, 640], conf_thres=0.25, iou_thres=0.45, max_det=1000, device=, view_img=False, save_txt=False, save_format=0, save_csv=False, save_conf=False, save_crop=False, nosave=False, classes=None, agnostic_nms=False, augment=False, visualize=False, update=False, project=runs/detect, name=exp, exist_ok=False, line_thickness=3, hide_labels=False, hide_conf=False, half=False, dnn=False, vid_stride=1 YOLOv5 π v7.0-378-g2f74455a Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB) Fusing layers... Model summary: 157 layers, 1761871 parameters, 0 gradients, 4.1 GFLOPs image 1/1 /kaggle/input/fireee/fire.109.jpg: 384x640 2 Fires, 27.7ms Speed: 0.5ms pre-process, 27.7ms inference, 134.5ms NMS per image at shape (1, 3, 640, 640) Results saved to runs/detect/exp3
Test on new imagesΒΆ
InΒ [Β ]:
import cv2
import matplotlib.pyplot as plt
# Path to the output image
output_image_path = "/kaggle/working/yolov5/runs/detect/exp3/fire.109.jpg"
# Load the output image
output_image = cv2.imread(output_image_path)
output_image = cv2.cvtColor(output_image, cv2.COLOR_BGR2RGB)
# Display the output image
plt.figure(figsize=(8, 8))
plt.imshow(output_image)
plt.title('Detected Image - fire.109.jpg')
plt.axis('off') # Hide axes
plt.show()
InΒ [Β ]:
# Running YOLOv5 prediction from the terminal
!python detect.py --weights /kaggle/working/yolov5/runs/train/exp/weights/best.pt --source /kaggle/input/fire-and-smoke-detection-for-yolov8/train/images/0058_jpg.rf.629dad210eae7f57976d164974befbe2.jpg --img 640 --conf 0.25
detect: weights=['/kaggle/working/yolov5/runs/train/exp/weights/best.pt'], source=/kaggle/input/fire-and-smoke-detection-for-yolov8/train/images/0058_jpg.rf.629dad210eae7f57976d164974befbe2.jpg, data=data/coco128.yaml, imgsz=[640, 640], conf_thres=0.25, iou_thres=0.45, max_det=1000, device=, view_img=False, save_txt=False, save_format=0, save_csv=False, save_conf=False, save_crop=False, nosave=False, classes=None, agnostic_nms=False, augment=False, visualize=False, update=False, project=runs/detect, name=exp, exist_ok=False, line_thickness=3, hide_labels=False, hide_conf=False, half=False, dnn=False, vid_stride=1 YOLOv5 π v7.0-378-g2f74455a Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB) Fusing layers... Model summary: 157 layers, 1761871 parameters, 0 gradients, 4.1 GFLOPs image 1/1 /kaggle/input/fire-and-smoke-detection-for-yolov8/train/images/0058_jpg.rf.629dad210eae7f57976d164974befbe2.jpg: 640x640 5 Fires, 1 Smoke, 6.2ms Speed: 0.6ms pre-process, 6.2ms inference, 133.8ms NMS per image at shape (1, 3, 640, 640) Results saved to runs/detect/exp4
InΒ [Β ]:
output_image_path = "/kaggle/working/yolov5/runs/detect/exp4/0058_jpg.rf.629dad210eae7f57976d164974befbe2.jpg" # Adjust if the file name is different
# Load the output image
output_image = cv2.imread(output_image_path)
output_image = cv2.cvtColor(output_image, cv2.COLOR_BGR2RGB)
# Display the output image
plt.figure(figsize=(8, 8))
plt.imshow(output_image)
plt.title('Detected Image - fire.109.jpg')
plt.axis('off') # Hide axes
plt.show()
YoloV7ΒΆ
InΒ [Β ]:
!pip install roboflow
from roboflow import Roboflow
rf = Roboflow(api_key="GZBlBPlRdRvB6uR7BqaS")
project = rf.workspace("detection-e83li").project("smokeandfire")
version = project.version(2)
dataset = version.download("yolov7")
InΒ [Β ]:
!git clone https://github.com/WongKinYiu/yolov7.git
%cd yolov7
Cloning into 'yolov7'... remote: Enumerating objects: 1197, done. remote: Total 1197 (delta 0), reused 0 (delta 0), pack-reused 1197 (from 1) Receiving objects: 100% (1197/1197), 74.23 MiB | 38.22 MiB/s, done. Resolving deltas: 100% (520/520), done. /kaggle/working/yolov7
InΒ [Β ]:
!pip install -r requirements.txt
InΒ [Β ]:
%env WANDB_MODE=disabled
env: WANDB_MODE=disabled
InΒ [Β ]:
import torch
torch.cuda.empty_cache()
InΒ [Β ]:
!python train.py --weights yolov7.pt --data '/kaggle/working/yolov7/smokeandfire-2/data.yaml' --epochs 80 --img-size 512 --batch-size 32
/kaggle/working/yolov7/train.py:71: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
run_id = torch.load(weights, map_location=device).get('wandb_id') if weights.endswith('.pt') and os.path.isfile(weights) else None
/kaggle/working/yolov7/train.py:87: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
ckpt = torch.load(weights, map_location=device) # load checkpoint
/opt/conda/lib/python3.10/site-packages/torch/functional.py:513: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/native/TensorShape.cpp:3609.)
return _VF.meshgrid(tensors, **kwargs) # type: ignore[attr-defined]
/kaggle/working/yolov7/utils/datasets.py:392: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
cache, exists = torch.load(cache_path), True # load
train: Scanning 'smokeandfire-2/train/labels.cache' images and labels... 8243 fo
/opt/conda/lib/python3.10/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock.
self.pid = os.fork()
val: Scanning 'smokeandfire-2/valid/labels.cache' images and labels... 1062 foun
autoanchor: Analyzing anchors... anchors/target = 4.83, Best Possible Recall (BPR) = 0.9982
/kaggle/working/yolov7/train.py:299: FutureWarning: `torch.cuda.amp.GradScaler(args...)` is deprecated. Please use `torch.amp.GradScaler('cuda', args...)` instead.
scaler = amp.GradScaler(enabled=cuda)
0%| | 0/258 [00:00<?, ?it/s]/kaggle/working/yolov7/train.py:360: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
with amp.autocast(enabled=cuda):
0/79 4.76G 0.06082 0.01174 0.01107 0.08362 93 512
Class Images Labels P R [email protected]
all 1062 2081 0.686 0.576 0.608 0.359
1/79 16G 0.04721 0.009793 0.004424 0.06143 96 512
Class Images Labels P R [email protected]
all 1062 2081 0.795 0.697 0.734 0.456
2/79 16.5G 0.04521 0.009044 0.003642 0.05789 106 512
Class Images Labels P R [email protected]
all 1062 2081 0.832 0.751 0.779 0.506
3/79 16.4G 0.04303 0.008948 0.003376 0.05536 107 512
Class Images Labels P R [email protected]
all 1062 2081 0.853 0.742 0.779 0.512
4/79 16.4G 0.04178 0.008852 0.003097 0.05373 90 512
Class Images Labels P R [email protected]
all 1062 2081 0.861 0.733 0.773 0.532
5/79 16.2G 0.0394 0.008794 0.002656 0.05085 104 512
Class Images Labels P R [email protected]
all 1062 2081 0.853 0.786 0.804 0.566
6/79 16.2G 0.0379 0.008572 0.002412 0.04888 82 512
Class Images Labels P R [email protected]
all 1062 2081 0.877 0.812 0.818 0.589
7/79 16.2G 0.0364 0.008379 0.002274 0.04706 120 512
Class Images Labels P R [email protected]
all 1062 2081 0.892 0.802 0.825 0.579
8/79 16.2G 0.03558 0.008295 0.002127 0.046 85 512
Class Images Labels P R [email protected]
all 1062 2081 0.873 0.798 0.813 0.589
9/79 16.2G 0.03499 0.008348 0.001946 0.04528 97 512
Class Images Labels P R [email protected]
all 1062 2081 0.881 0.817 0.823 0.6
10/79 16.2G 0.0346 0.008206 0.001891 0.0447 95 512
Class Images Labels P R [email protected]
all 1062 2081 0.89 0.821 0.827 0.61
11/79 16.2G 0.03381 0.008055 0.00184 0.04371 120 512
Class Images Labels P R [email protected]
all 1062 2081 0.891 0.813 0.836 0.618
12/79 16.2G 0.03386 0.008203 0.001912 0.04397 98 512
Class Images Labels P R [email protected]
all 1062 2081 0.88 0.816 0.811 0.612
13/79 16.6G 0.03323 0.008034 0.001781 0.04304 119 512
Class Images Labels P R [email protected]
all 1062 2081 0.894 0.829 0.842 0.628
14/79 16.6G 0.03283 0.007888 0.001682 0.0424 125 512
Class Images Labels P R [email protected]
all 1062 2081 0.88 0.837 0.838 0.615
15/79 16.6G 0.03255 0.007959 0.001786 0.04229 106 512
Class Images Labels P R [email protected]
all 1062 2081 0.894 0.84 0.839 0.637
16/79 16.6G 0.03191 0.00791 0.001592 0.04141 77 512
Class Images Labels P R [email protected]
all 1062 2081 0.888 0.832 0.834 0.637
17/79 16.6G 0.03207 0.007965 0.001587 0.04162 103 512
Class Images Labels P R [email protected]
all 1062 2081 0.887 0.832 0.836 0.638
18/79 16.6G 0.03184 0.007968 0.001629 0.04143 102 512
Class Images Labels P R [email protected]
all 1062 2081 0.887 0.84 0.842 0.635
19/79 16.6G 0.03177 0.007939 0.001539 0.04125 105 512
Class Images Labels P R [email protected]
all 1062 2081 0.879 0.836 0.827 0.636
20/79 16.6G 0.0313 0.007789 0.001449 0.04053 86 512
Class Images Labels P R [email protected]
all 1062 2081 0.902 0.842 0.853 0.66
21/79 16.6G 0.03113 0.007776 0.001419 0.04032 79 512
Class Images Labels P R [email protected]
all 1062 2081 0.884 0.845 0.834 0.637
22/79 16.6G 0.03101 0.007646 0.001505 0.04016 102 512
Class Images Labels P R [email protected]
all 1062 2081 0.894 0.835 0.838 0.645
23/79 16.6G 0.03116 0.007816 0.001545 0.04052 77 512
Class Images Labels P R [email protected]
all 1062 2081 0.864 0.854 0.841 0.647
24/79 16.6G 0.03057 0.007738 0.001449 0.03976 113 512
Class Images Labels P R [email protected]
all 1062 2081 0.869 0.844 0.835 0.642
25/79 16.6G 0.03045 0.007681 0.001393 0.03953 76 512
Class Images Labels P R [email protected]
all 1062 2081 0.885 0.844 0.836 0.652
26/79 16.6G 0.03031 0.007759 0.001401 0.03947 75 512
Class Images Labels P R [email protected]
all 1062 2081 0.882 0.848 0.841 0.655
27/79 16.6G 0.03039 0.007828 0.001375 0.03959 73 512
Class Images Labels P R [email protected]
all 1062 2081 0.895 0.832 0.849 0.663
28/79 16.6G 0.03013 0.007641 0.001332 0.0391 102 512
Class Images Labels P R [email protected]
all 1062 2081 0.882 0.853 0.844 0.659
29/79 16.6G 0.02987 0.00765 0.001274 0.03879 92 512
Class Images Labels P R [email protected]
all 1062 2081 0.896 0.841 0.845 0.66
30/79 16.6G 0.02957 0.007621 0.00126 0.03845 127 512
Class Images Labels P R [email protected]
all 1062 2081 0.894 0.849 0.852 0.669
31/79 16.6G 0.02985 0.007665 0.001287 0.0388 89 512
Class Images Labels P R [email protected]
all 1062 2081 0.898 0.84 0.848 0.666
33/79 16.6G 0.0291 0.007554 0.001196 0.03785 120 512
Class Images Labels P R [email protected]
all 1062 2081 0.898 0.846 0.851 0.668
34/79 16.6G 0.02895 0.007451 0.001278 0.03768 93 512
Class Images Labels P R [email protected]
all 1062 2081 0.884 0.86 0.856 0.673
35/79 16.6G 0.02845 0.007434 0.001199 0.03709 127 512
Class Images Labels P R [email protected]
all 1062 2081 0.898 0.853 0.861 0.682
36/79 16.6G 0.02875 0.007513 0.001189 0.03745 101 512
Class Images Labels P R [email protected]
all 1062 2081 0.895 0.854 0.855 0.677
37/79 16.6G 0.02829 0.00742 0.001157 0.03686 102 512
Class Images Labels P R [email protected]
all 1062 2081 0.899 0.868 0.86 0.679
38/79 16.6G 0.02828 0.007494 0.001138 0.03691 106 512
Class Images Labels P R [email protected]
all 1062 2081 0.895 0.859 0.86 0.684
39/79 16.6G 0.02822 0.00737 0.001184 0.03678 115 512
Class Images Labels P R [email protected]
all 1062 2081 0.896 0.846 0.849 0.676
40/79 16.6G 0.0282 0.007337 0.001047 0.03659 83 512
Class Images Labels P R [email protected]
all 1062 2081 0.893 0.863 0.853 0.682
41/79 16.6G 0.02778 0.007388 0.001102 0.03627 83 512
Class Images Labels P R [email protected]
all 1062 2081 0.903 0.864 0.855 0.68
42/79 16.6G 0.02723 0.007376 0.001075 0.03568 97 512
Class Images Labels P R [email protected]
all 1062 2081 0.901 0.859 0.863 0.686
43/79 16.6G 0.02751 0.007377 0.000997 0.03588 99 512
Class Images Labels P R [email protected]
all 1062 2081 0.887 0.874 0.867 0.688
44/79 16.6G 0.02759 0.007126 0.001071 0.03579 87 512
Class Images Labels P R [email protected]
all 1062 2081 0.902 0.855 0.867 0.691
45/79 16.6G 0.02713 0.007168 0.001091 0.03539 92 512
Class Images Labels P R [email protected]
all 1062 2081 0.895 0.863 0.855 0.685
46/79 16.6G 0.02688 0.007018 0.001025 0.03492 110 512
Class Images Labels P R [email protected]
all 1062 2081 0.9 0.855 0.865 0.695
47/79 16.6G 0.02653 0.007125 0.0009582 0.03461 86 512
48/79 16.6G 0.02627 0.00703 0.0009608 0.03426 97 512
Class Images Labels P R [email protected]
all 1062 2081 0.903 0.864 0.864 0.691
49/79 16.6G 0.02663 0.007066 0.000989 0.03468 118 512
Class Images Labels P R [email protected]
all 1062 2081 0.904 0.858 0.869 0.695
50/79 16.6G 0.02589 0.007052 0.0009219 0.03386 83 512
Class Images Labels P R [email protected]
all 1062 2081 0.902 0.868 0.869 0.695
51/79 16.6G 0.0261 0.006989 0.0009679 0.03406 86 512
Class Images Labels P R [email protected]
all 1062 2081 0.91 0.862 0.87 0.697
52/79 16.6G 0.02608 0.007004 0.0009319 0.03402 81 512
Class Images Labels P R [email protected]
all 1062 2081 0.906 0.866 0.87 0.7
53/79 16.6G 0.02575 0.006857 0.0009326 0.03354 103 512
Class Images Labels P R [email protected]
all 1062 2081 0.905 0.866 0.873 0.701
54/79 16.6G 0.02555 0.006881 0.0008876 0.03332 110 512
Class Images Labels P R [email protected]
all 1062 2081 0.91 0.86 0.866 0.697
55/79 16.6G 0.0256 0.00687 0.0008233 0.0333 108 512
Class Images Labels P R [email protected]
all 1062 2081 0.9 0.867 0.875 0.704
56/79 16.6G 0.02556 0.006961 0.0008927 0.03342 70 512
Class Images Labels P R [email protected]
all 1062 2081 0.9 0.866 0.875 0.706
57/79 16.6G 0.02504 0.006787 0.0007799 0.03261 63 512
Class Images Labels P R [email protected]
all 1062 2081 0.906 0.867 0.874 0.709
58/79 16.6G 0.02496 0.006737 0.0008432 0.03254 110 512
Class Images Labels P R [email protected]
all 1062 2081 0.9 0.872 0.873 0.706
59/79 16.6G 0.02496 0.006685 0.0008593 0.03251 76 512
Class Images Labels P R [email protected]
all 1062 2081 0.909 0.87 0.875 0.708
60/79 16.6G 0.02458 0.006748 0.0008318 0.03216 67 512
Class Images Labels P R [email protected]
all 1062 2081 0.911 0.864 0.874 0.711
61/79 16.6G 0.02452 0.006734 0.0007902 0.03205 122 512
Class Images Labels P R [email protected]
all 1062 2081 0.909 0.868 0.875 0.711
62/79 16.6G 0.02422 0.006549 0.0007748 0.03155 101 512
Class Images Labels P R [email protected]
all 1062 2081 0.901 0.868 0.878 0.714
63/79 16.6G 0.02407 0.006659 0.0008269 0.03155 121 512
Class Images Labels P R [email protected]
all 1062 2081 0.903 0.862 0.876 0.709
64/79 16.6G 0.02394 0.006484 0.0008308 0.03125 88 512
Class Images Labels P R [email protected]
all 1062 2081 0.908 0.868 0.877 0.714
65/79 16.6G 0.02387 0.00651 0.0008168 0.0312 99 512
Class Images Labels P R [email protected]
all 1062 2081 0.909 0.867 0.88 0.714
66/79 16.6G 0.0238 0.006468 0.0007055 0.03098 76 512
Class Images Labels P R [email protected]
all 1062 2081 0.912 0.865 0.882 0.715
67/79 16.6G 0.02364 0.006409 0.0007594 0.03081 91 512
Class Images Labels P R [email protected]
all 1062 2081 0.915 0.865 0.882 0.715
68/79 16.6G 0.02325 0.006264 0.0007115 0.03022 155 512
Class Images Labels P R [email protected]
all 1062 2081 0.914 0.862 0.882 0.717
69/79 16.6G 0.02338 0.006446 0.0007162 0.03054 113 512
Class Images Labels P R [email protected]
all 1062 2081 0.897 0.874 0.88 0.717
70/79 16.6G 0.02324 0.006347 0.0006713 0.03026 119 512
Class Images Labels P R [email protected]
all 1062 2081 0.9 0.872 0.879 0.715
71/79 16.6G 0.02341 0.006388 0.0007474 0.03055 101 512
Class Images Labels P R [email protected]
all 1062 2081 0.895 0.872 0.88 0.713
72/79 16.6G 0.02285 0.006408 0.0006716 0.02993 106 512
Class Images Labels P R [email protected]
all 1062 2081 0.901 0.873 0.88 0.714
73/79 16.6G 0.02268 0.006182 0.0007327 0.02959 96 512
Class Images Labels P R [email protected]
all 1062 2081 0.909 0.865 0.88 0.713
74/79 16.6G 0.02253 0.006277 0.0006819 0.02949 114 512
Class Images Labels P R [email protected]
all 1062 2081 0.909 0.869 0.881 0.716
75/79 16.6G 0.02274 0.006218 0.0006566 0.02962 123 512
Class Images Labels P R [email protected]
all 1062 2081 0.908 0.868 0.883 0.717
76/79 16.6G 0.02271 0.006204 0.0006443 0.02956 74 512
Class Images Labels P R [email protected]
all 1062 2081 0.897 0.875 0.883 0.717
77/79 16.6G 0.02273 0.006286 0.0006503 0.02966 98 512
Class Images Labels P R [email protected]
all 1062 2081 0.897 0.872 0.883 0.717
78/79 16.6G 0.02258 0.006268 0.0006554 0.02951 65 512
Class Images Labels P R [email protected]
all 1062 2081 0.899 0.872 0.882 0.72
79/79 16.6G 0.02245 0.006137 0.0006392 0.02923 110 512
Class Images Labels P R [email protected]
all 1062 2081 0.905 0.87 0.882 0.72
Fire 1062 1320 0.917 0.942 0.941 0.779
Smoke 1062 761 0.893 0.797 0.822 0.66
/kaggle/working/yolov7/utils/general.py:802: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
x = torch.load(f, map_location=torch.device('cpu'))
Optimizer stripped from runs/train/exp8/weights/last.pt, 74.8MB
Optimizer stripped from runs/train/exp8/weights/best.pt, 74.8MB
InΒ [Β ]:
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
img = mpimg.imread('/kaggle/working/yolov7/runs/train/exp8/confusion_matrix.png')
imgplot = plt.imshow(img)
plt.show()
InΒ [Β ]:
img = mpimg.imread('/kaggle/working/yolov7/runs/train/exp8/P_curve.png')
imgplot = plt.imshow(img)
plt.show()
InΒ [Β ]:
!python detect.py --weights /kaggle/working/yolov7/runs/train/exp8/weights/best.pt --img 640 --conf 0.35 --source /kaggle/working/yolov7/smokeandfire-2/test/images
InΒ [Β ]:
!python detect.py --weights /kaggle/working/yolov7/runs/train/exp8/weights/best.pt --source /kaggle/input/fireee/fire.109.jpg --img 640 --conf 0.25
test on new imagesΒΆ
InΒ [Β ]:
output_image_path = "/kaggle/working/yolov7/runs/detect/exp3/fire.109.jpg"
output_image = cv2.imread(output_image_path)
output_image = cv2.cvtColor(output_image, cv2.COLOR_BGR2RGB)
plt.figure(figsize=(8, 8))
plt.imshow(output_image)
plt.title('Detected Image - fire.109.jpg')
plt.axis('off')
plt.show()
InΒ [Β ]:
# Running YOLOv5 prediction
!python detect.py --weights /kaggle/working/yolov7/runs/train/exp8/weights/best.pt --source /kaggle/input/fire-and-smoke-detection-for-yolov8/train/images/0058_jpg.rf.629dad210eae7f57976d164974befbe2.jpg --img 640 --conf 0.25
InΒ [Β ]:
output_image_path = "/kaggle/working/yolov7/runs/detect/exp4/0058_jpg.rf.629dad210eae7f57976d164974befbe2.jpg"
output_image = cv2.imread(output_image_path)
output_image = cv2.cvtColor(output_image, cv2.COLOR_BGR2RGB)
plt.figure(figsize=(8, 8))
plt.imshow(output_image)
plt.title('Detected Image - fire.109.jpg')
plt.axis('off')
plt.show()
yolo v7 best but its more complex
yolo v8-enhancementΒΆ
InΒ [Β ]:
!yolo train model=yolov8s.pt data='/kaggle/working/smokeandfire-2/data.yaml' epochs=120 imgsz=640 batch=32 augment=True lr0=0.005 lrf=0.01 weight_decay=0.0005 optimizer=SGD mosaic=True hsv_h=0.015 hsv_s=0.7 hsv_v=0.4 degrees=15 translate=0.1 scale=0.5 shear=0.2 flipud=0.5 fliplr=0.5 mixup=0.2
Ultralytics 8.3.48 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB) engine/trainer: task=detect, mode=train, model=yolov8s.pt, data=/kaggle/working/smokeandfire-2/data.yaml, epochs=120, time=None, patience=100, batch=32, imgsz=640, save=True, save_period=-1, cache=False, device=None, workers=8, project=None, name=train2, exist_ok=False, pretrained=True, optimizer=SGD, verbose=True, seed=0, deterministic=True, single_cls=False, rect=False, cos_lr=False, close_mosaic=10, resume=False, amp=True, fraction=1.0, profile=False, freeze=None, multi_scale=False, overlap_mask=True, mask_ratio=4, dropout=0.0, val=True, split=val, save_json=False, save_hybrid=False, conf=None, iou=0.7, max_det=300, half=False, dnn=False, plots=True, source=None, vid_stride=1, stream_buffer=False, visualize=False, augment=True, agnostic_nms=False, classes=None, retina_masks=False, embed=None, show=False, save_frames=False, save_txt=False, save_conf=False, save_crop=False, show_labels=True, show_conf=True, show_boxes=True, line_width=None, format=torchscript, keras=False, optimize=False, int8=False, dynamic=False, simplify=True, opset=None, workspace=None, nms=False, lr0=0.005, lrf=0.01, momentum=0.937, weight_decay=0.0005, warmup_epochs=3.0, warmup_momentum=0.8, warmup_bias_lr=0.1, box=7.5, cls=0.5, dfl=1.5, pose=12.0, kobj=1.0, nbs=64, hsv_h=0.015, hsv_s=0.7, hsv_v=0.4, degrees=15, translate=0.1, scale=0.5, shear=0.2, perspective=0.0, flipud=0.5, fliplr=0.5, bgr=0.0, mosaic=True, mixup=0.2, copy_paste=0.0, copy_paste_mode=flip, auto_augment=randaugment, erasing=0.4, crop_fraction=1.0, cfg=None, tracker=botsort.yaml, save_dir=runs/detect/train2 Overriding model.yaml nc=80 with nc=2 from n params module arguments 0 -1 1 928 ultralytics.nn.modules.conv.Conv [3, 32, 3, 2] 1 -1 1 18560 ultralytics.nn.modules.conv.Conv [32, 64, 3, 2] 2 -1 1 29056 ultralytics.nn.modules.block.C2f [64, 64, 1, True] 3 -1 1 73984 ultralytics.nn.modules.conv.Conv [64, 128, 3, 2] 4 -1 2 197632 ultralytics.nn.modules.block.C2f [128, 128, 2, True] 5 -1 1 295424 ultralytics.nn.modules.conv.Conv [128, 256, 3, 2] 6 -1 2 788480 ultralytics.nn.modules.block.C2f [256, 256, 2, True] 7 -1 1 1180672 ultralytics.nn.modules.conv.Conv [256, 512, 3, 2] 8 -1 1 1838080 ultralytics.nn.modules.block.C2f [512, 512, 1, True] 9 -1 1 656896 ultralytics.nn.modules.block.SPPF [512, 512, 5] 10 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 11 [-1, 6] 1 0 ultralytics.nn.modules.conv.Concat [1] 12 -1 1 591360 ultralytics.nn.modules.block.C2f [768, 256, 1] 13 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 14 [-1, 4] 1 0 ultralytics.nn.modules.conv.Concat [1] 15 -1 1 148224 ultralytics.nn.modules.block.C2f [384, 128, 1] 16 -1 1 147712 ultralytics.nn.modules.conv.Conv [128, 128, 3, 2] 17 [-1, 12] 1 0 ultralytics.nn.modules.conv.Concat [1] 18 -1 1 493056 ultralytics.nn.modules.block.C2f [384, 256, 1] 19 -1 1 590336 ultralytics.nn.modules.conv.Conv [256, 256, 3, 2] 20 [-1, 9] 1 0 ultralytics.nn.modules.conv.Concat [1] 21 -1 1 1969152 ultralytics.nn.modules.block.C2f [768, 512, 1] 22 [15, 18, 21] 1 2116822 ultralytics.nn.modules.head.Detect [2, [128, 256, 512]] Model summary: 225 layers, 11,136,374 parameters, 11,136,358 gradients, 28.6 GFLOPs Transferred 349/355 items from pretrained weights TensorBoard: Start with 'tensorboard --logdir runs/detect/train2', view at http://localhost:6006/ Freezing layer 'model.22.dfl.conv.weight' AMP: running Automatic Mixed Precision (AMP) checks... AMP: checks passed β train: Scanning /kaggle/working/smokeandfire-2/train/labels.cache... 8243 images WARNING β οΈ Box and segment counts should be equal, but got len(segments) = 12, len(boxes) = 16365. To resolve this only boxes will be used and all segments will be removed. To avoid this please supply either a detect or segment dataset, not a detect-segment mixed dataset. /opt/conda/lib/python3.10/site-packages/albumentations/__init__.py:24: UserWarning: A new version of Albumentations is available: 1.4.22 (you have 1.4.21). Upgrade using: pip install -U albumentations. To disable automatic update checks, set the environment variable NO_ALBUMENTATIONS_UPDATE to 1. check_for_updates() albumentations: Blur(p=0.01, blur_limit=(3, 7)), MedianBlur(p=0.01, blur_limit=(3, 7)), ToGray(p=0.01, num_output_channels=3, method='weighted_average'), CLAHE(p=0.01, clip_limit=(1.0, 4.0), tile_grid_size=(8, 8)) /opt/conda/lib/python3.10/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock. self.pid = os.fork() val: Scanning /kaggle/working/smokeandfire-2/valid/labels.cache... 1062 images, Plotting labels to runs/detect/train2/labels.jpg... optimizer: SGD(lr=0.005, momentum=0.937) with parameter groups 57 weight(decay=0.0), 64 weight(decay=0.0005), 63 bias(decay=0.0) TensorBoard: model graph visualization added β Image sizes 640 train, 640 val Using 4 dataloader workers Logging results to runs/detect/train2 Starting training for 120 epochs... Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 1/120 8.04G 1.772 2.504 1.853 96 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.729 0.653 0.667 0.467 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 2/120 8.24G 1.449 1.591 1.59 76 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.822 0.746 0.783 0.552 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 3/120 8.29G 1.451 1.499 1.583 53 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.748 0.698 0.727 0.517 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 4/120 8.29G 1.473 1.54 1.608 90 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.834 0.7 0.75 0.503 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 5/120 8.29G 1.459 1.493 1.58 95 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.801 0.678 0.727 0.517 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 6/120 8.37G 1.424 1.442 1.568 62 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.849 0.776 0.8 0.571 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 7/120 8.38G 1.401 1.399 1.542 72 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.835 0.771 0.802 0.568 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 8/120 8.28G 1.39 1.377 1.545 82 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.843 0.78 0.811 0.581 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 9/120 8.28G 1.366 1.36 1.531 54 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.853 0.797 0.812 0.591 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 10/120 8.28G 1.352 1.331 1.516 91 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.844 0.813 0.822 0.602 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 11/120 8.28G 1.339 1.305 1.509 88 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.859 0.813 0.82 0.597 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 12/120 8.3G 1.326 1.287 1.496 112 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.88 0.8 0.823 0.611 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 13/120 8.3G 1.318 1.273 1.486 76 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.862 0.808 0.813 0.6 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 14/120 8.28G 1.312 1.26 1.488 80 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.85 0.808 0.828 0.618 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 15/120 8.36G 1.306 1.268 1.487 114 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.869 0.811 0.827 0.624 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 16/120 8.36G 1.284 1.238 1.477 91 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.878 0.833 0.837 0.619 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 17/120 8.36G 1.28 1.223 1.466 65 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.874 0.832 0.843 0.636 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 18/120 8.36G 1.277 1.224 1.472 137 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.862 0.818 0.83 0.627 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 19/120 8.37G 1.279 1.208 1.474 92 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.873 0.826 0.842 0.641 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 20/120 8.28G 1.273 1.205 1.464 92 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.877 0.835 0.851 0.648 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 21/120 8.29G 1.257 1.176 1.45 75 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.89 0.826 0.843 0.639 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 22/120 8.37G 1.266 1.183 1.45 87 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.876 0.837 0.845 0.648 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 23/120 8.37G 1.251 1.171 1.45 117 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.881 0.827 0.848 0.642 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 24/120 8.37G 1.243 1.165 1.436 129 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.874 0.839 0.854 0.66 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 25/120 8.29G 1.235 1.144 1.428 80 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.892 0.842 0.857 0.653 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 26/120 8.3G 1.24 1.145 1.424 78 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.876 0.84 0.856 0.653 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 27/120 8.37G 1.22 1.136 1.423 75 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.879 0.833 0.851 0.653 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 28/120 8.29G 1.23 1.132 1.431 71 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.883 0.846 0.86 0.662 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 29/120 8.29G 1.22 1.119 1.416 71 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.893 0.85 0.863 0.67 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 30/120 8.3G 1.215 1.109 1.41 70 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.88 0.855 0.86 0.67 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 31/120 8.38G 1.199 1.094 1.405 68 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.884 0.847 0.853 0.667 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 32/120 8.36G 1.205 1.104 1.405 75 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.89 0.846 0.857 0.668 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 33/120 8.37G 1.204 1.098 1.401 97 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.844 0.863 0.67 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 34/120 8.29G 1.203 1.087 1.405 72 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.898 0.843 0.863 0.673 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 35/120 8.29G 1.19 1.091 1.406 72 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.886 0.845 0.863 0.675 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 36/120 8.37G 1.171 1.061 1.387 100 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.887 0.847 0.861 0.68 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 37/120 8.29G 1.182 1.057 1.388 81 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.893 0.845 0.859 0.676 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 38/120 8.3G 1.19 1.073 1.398 74 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.898 0.844 0.863 0.678 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 39/120 8.3G 1.176 1.05 1.386 98 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.843 0.86 0.678 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 40/120 8.29G 1.164 1.033 1.38 77 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.903 0.844 0.86 0.675 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 41/120 8.37G 1.176 1.047 1.384 78 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.893 0.853 0.86 0.688 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 42/120 8.29G 1.166 1.035 1.374 111 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.85 0.861 0.686 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 43/120 8.29G 1.175 1.036 1.375 111 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.896 0.85 0.86 0.682 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 44/120 8.29G 1.151 1.017 1.369 67 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.894 0.85 0.865 0.688 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 45/120 8.28G 1.157 1.022 1.37 79 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.901 0.85 0.865 0.687 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 46/120 8.28G 1.153 1.004 1.363 94 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.891 0.85 0.867 0.689 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 47/120 8.3G 1.136 1.002 1.354 77 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.9 0.855 0.866 0.69 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 48/120 8.29G 1.158 1.018 1.371 57 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.903 0.847 0.865 0.688 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 49/120 8.29G 1.14 0.9994 1.355 96 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.9 0.856 0.871 0.693 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 50/120 8.29G 1.139 0.9899 1.358 91 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.893 0.856 0.865 0.691 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 51/120 8.37G 1.135 0.9923 1.353 76 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.901 0.864 0.867 0.693 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 52/120 8.29G 1.138 0.9864 1.356 99 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.906 0.856 0.869 0.693 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 53/120 8.37G 1.11 0.9659 1.333 103 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.905 0.853 0.867 0.697 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 54/120 8.37G 1.122 0.9754 1.344 100 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.906 0.858 0.866 0.697 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 55/120 8.29G 1.122 0.9694 1.347 78 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.91 0.858 0.87 0.7 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 56/120 8.29G 1.115 0.964 1.34 68 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.904 0.862 0.868 0.701 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 57/120 8.38G 1.115 0.9537 1.333 76 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.904 0.865 0.868 0.703 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 58/120 8.3G 1.11 0.9469 1.331 61 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.907 0.855 0.868 0.697 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 59/120 8.37G 1.111 0.9461 1.333 108 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.911 0.859 0.864 0.698 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 60/120 8.37G 1.125 0.9612 1.34 117 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.908 0.862 0.874 0.704 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 61/120 8.37G 1.105 0.9402 1.327 69 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.898 0.866 0.873 0.705 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 62/120 8.29G 1.11 0.9479 1.329 66 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.901 0.866 0.871 0.703 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 63/120 8.37G 1.093 0.9297 1.32 80 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.905 0.867 0.874 0.703 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 64/120 8.37G 1.088 0.9111 1.31 58 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.903 0.864 0.872 0.699 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 65/120 8.3G 1.088 0.9161 1.31 86 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.906 0.86 0.872 0.705 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 66/120 8.37G 1.088 0.9072 1.31 128 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.914 0.861 0.869 0.702 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 67/120 8.37G 1.064 0.8933 1.298 53 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.914 0.868 0.872 0.703 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 68/120 8.37G 1.09 0.9137 1.316 97 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.907 0.865 0.872 0.703 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 69/120 8.29G 1.072 0.8963 1.301 98 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.91 0.859 0.874 0.706 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 70/120 8.3G 1.073 0.8986 1.307 68 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.906 0.862 0.872 0.701 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 71/120 8.29G 1.073 0.8938 1.302 71 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.902 0.868 0.872 0.704 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 72/120 8.38G 1.079 0.8943 1.312 75 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.869 0.873 0.705 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 73/120 8.37G 1.069 0.8782 1.302 67 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.902 0.864 0.871 0.704 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 74/120 8.3G 1.07 0.8827 1.297 73 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.902 0.861 0.871 0.703 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 75/120 8.38G 1.065 0.8809 1.295 84 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.91 0.859 0.871 0.706 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 76/120 8.37G 1.054 0.8663 1.284 87 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.915 0.854 0.872 0.708 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 77/120 8.37G 1.059 0.8701 1.294 75 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.913 0.862 0.875 0.708 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 78/120 7.91G 1.055 0.8605 1.287 102 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.91 0.856 0.874 0.707 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 79/120 8.3G 1.045 0.8522 1.283 101 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.914 0.853 0.872 0.706 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 80/120 8.3G 1.041 0.8417 1.279 101 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.906 0.856 0.872 0.706 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 81/120 8.37G 1.037 0.8555 1.286 94 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.908 0.859 0.873 0.709 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 82/120 8.37G 1.052 0.8587 1.284 87 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.904 0.859 0.873 0.708 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 83/120 8.29G 1.034 0.8497 1.274 66 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.906 0.858 0.873 0.709 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 84/120 8.3G 1.043 0.8464 1.283 95 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.902 0.861 0.873 0.709 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 85/120 8.37G 1.037 0.8386 1.276 73 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.893 0.866 0.873 0.71 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 86/120 8.28G 1.036 0.8349 1.277 71 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.9 0.863 0.874 0.71 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 87/120 8.3G 1.029 0.8345 1.271 80 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.898 0.861 0.873 0.711 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 88/120 8.3G 1.02 0.8315 1.269 94 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.906 0.861 0.874 0.71 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 89/120 8.29G 1.011 0.8078 1.259 57 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.901 0.865 0.874 0.71 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 90/120 8.3G 1.021 0.8177 1.259 91 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.901 0.863 0.874 0.712 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 91/120 8.37G 1.007 0.8019 1.256 53 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.902 0.862 0.874 0.713 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 92/120 8.37G 1.014 0.8053 1.263 105 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.903 0.86 0.875 0.713 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 93/120 7.92G 1.019 0.817 1.261 103 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.9 0.865 0.876 0.715 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 94/120 8.29G 0.9994 0.7937 1.245 56 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.896 0.868 0.875 0.713 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 95/120 8.29G 1.006 0.8044 1.248 77 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.897 0.867 0.876 0.712 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 96/120 8.29G 1.013 0.809 1.255 122 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.896 0.869 0.877 0.712 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 97/120 8.37G 0.9935 0.7794 1.242 91 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.896 0.866 0.876 0.712 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 98/120 8.3G 0.9931 0.782 1.243 87 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.897 0.868 0.876 0.711 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 99/120 7.91G 0.9949 0.7824 1.242 72 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.894 0.866 0.876 0.713 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 100/120 8.28G 0.9929 0.7789 1.241 61 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.866 0.877 0.712 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 101/120 8.37G 0.9967 0.7906 1.248 90 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.894 0.866 0.876 0.712 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 102/120 8.28G 0.9856 0.7765 1.239 88 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.863 0.875 0.711 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 103/120 8.28G 0.98 0.7715 1.24 59 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.859 0.875 0.712 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 104/120 8.28G 0.9682 0.7546 1.231 93 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.858 0.875 0.712 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 105/120 8.35G 0.982 0.7707 1.239 111 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.861 0.876 0.713 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 106/120 8.37G 0.9774 0.7621 1.235 96 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.863 0.876 0.713 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 107/120 8.35G 0.9844 0.7555 1.239 73 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.861 0.875 0.713 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 108/120 8.28G 0.9711 0.7594 1.234 85 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.897 0.86 0.875 0.714 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 109/120 8.28G 0.9674 0.7519 1.226 78 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.894 0.862 0.875 0.714 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 110/120 8.39G 0.9586 0.7433 1.225 67 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.894 0.862 0.875 0.713 Closing dataloader mosaic albumentations: Blur(p=0.01, blur_limit=(3, 7)), MedianBlur(p=0.01, blur_limit=(3, 7)), ToGray(p=0.01, num_output_channels=3, method='weighted_average'), CLAHE(p=0.01, clip_limit=(1.0, 4.0), tile_grid_size=(8, 8)) /opt/conda/lib/python3.10/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock. self.pid = os.fork() Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 111/120 8.36G 0.8537 0.5345 1.137 44 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.897 0.863 0.875 0.712 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 112/120 8.28G 0.8322 0.5039 1.123 27 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.896 0.862 0.875 0.712 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 113/120 8.36G 0.8322 0.4976 1.124 34 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.897 0.862 0.875 0.712 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 114/120 8.28G 0.8217 0.4903 1.117 34 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.897 0.86 0.875 0.713 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 115/120 8.35G 0.8126 0.4803 1.112 41 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.896 0.861 0.875 0.713 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 116/120 8.37G 0.8127 0.4726 1.112 33 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.861 0.875 0.713 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 117/120 8.28G 0.8041 0.4765 1.105 35 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.896 0.859 0.875 0.713 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 118/120 8.29G 0.8052 0.4723 1.105 43 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.859 0.876 0.714 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 119/120 8.29G 0.7986 0.4622 1.101 39 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.896 0.86 0.875 0.714 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 120/120 8.29G 0.7987 0.461 1.105 35 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.896 0.86 0.875 0.715 120 epochs completed in 4.772 hours. Optimizer stripped from runs/detect/train2/weights/last.pt, 22.5MB Optimizer stripped from runs/detect/train2/weights/best.pt, 22.5MB Validating runs/detect/train2/weights/best.pt... Ultralytics 8.3.48 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB) Model summary (fused): 168 layers, 11,126,358 parameters, 0 gradients, 28.4 GFLOPs Class Images Instances Box(P R mAP50 m all 1062 2081 0.898 0.862 0.879 0.718 Fire 859 1320 0.911 0.938 0.939 0.765 Smoke 559 761 0.885 0.787 0.819 0.67 Speed: 0.1ms preprocess, 10.0ms inference, 0.0ms loss, 0.7ms postprocess per image Results saved to runs/detect/train2 π‘ Learn more at https://docs.ultralytics.com/modes/train
InΒ [Β ]:
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
img = mpimg.imread('/kaggle/working/runs/detect/train2/results.png')
imgplot = plt.imshow(img)
plt.show()
InΒ [Β ]:
img = mpimg.imread('/kaggle/working/runs/detect/train2/P_curve.png')
imgplot = plt.imshow(img)
plt.show()
InΒ [Β ]:
!yolo task=detect mode=predict model="/kaggle/working/runs/detect/train2/weights/best.pt" device =0 source="/kaggle/working/smokeandfire-2/test/images"
InΒ [Β ]:
!yolo predict model="/kaggle/working/runs/detect/train2/weights/best.pt", source="/kaggle/input/fire-and-smoke-detection-for-yolov8/test/images/40_jpg.rf.61555964ef797e1c3989f692641d30f7.jpg"
WARNING β οΈ argument 'model=/kaggle/working/runs/detect/train2/weights/best.pt,' does not require trailing comma ',', updating to 'model=/kaggle/working/runs/detect/train2/weights/best.pt'.
Ultralytics 8.3.48 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB)
Model summary (fused): 168 layers, 11,126,358 parameters, 0 gradients, 28.4 GFLOPs
image 1/1 /kaggle/input/fire-and-smoke-detection-for-yolov8/test/images/40_jpg.rf.61555964ef797e1c3989f692641d30f7.jpg: 640x640 1 Smoke, 8.5ms
Speed: 6.6ms preprocess, 8.5ms inference, 164.9ms postprocess per image at shape (1, 3, 640, 640)
Results saved to runs/detect/predict4
π‘ Learn more at https://docs.ultralytics.com/modes/predict
InΒ [Β ]:
output_image_path = "/kaggle/working/runs/detect/predict4/40_jpg.rf.61555964ef797e1c3989f692641d30f7.jpg"
output_image = cv2.imread(output_image_path)
output_image = cv2.cvtColor(output_image, cv2.COLOR_BGR2RGB)
# Display the output image
plt.figure(figsize=(8, 8))
plt.imshow(output_image)
plt.title('Detected Image - fire.109.jpg')
plt.axis('off')
plt.show()
InΒ [Β ]:
!yolo predict model="/kaggle/working/runs/detect/train2/weights/best.pt", source="/kaggle/input/fireimage/fire.109.jpg"
WARNING β οΈ argument 'model=/kaggle/working/runs/detect/train2/weights/best.pt,' does not require trailing comma ',', updating to 'model=/kaggle/working/runs/detect/train2/weights/best.pt'.
Ultralytics 8.3.48 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB)
Model summary (fused): 168 layers, 11,126,358 parameters, 0 gradients, 28.4 GFLOPs
image 1/1 /kaggle/input/fireimage/fire.109.jpg: 384x640 2 Fires, 51.7ms
Speed: 5.1ms preprocess, 51.7ms inference, 155.0ms postprocess per image at shape (1, 3, 384, 640)
Results saved to runs/detect/predict5
π‘ Learn more at https://docs.ultralytics.com/modes/predict
InΒ [Β ]:
output_image_path = "/kaggle/working/runs/detect/predict5/fire.109.jpg"
output_image = cv2.imread(output_image_path)
output_image = cv2.cvtColor(output_image, cv2.COLOR_BGR2RGB)
plt.figure(figsize=(8, 8))
plt.imshow(output_image)
plt.title('Detected Image - fire.109.jpg')
plt.axis('off')
plt.show()
yolo v11ΒΆ
InΒ [3]:
!yolo train model=yolo11n.pt data='/kaggle/working/smokeandfire-2/data.yaml' epochs=120 imgsz=512 batch=32 augment=True
Downloading https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11n.pt to 'yolo11n.pt'... 100%|ββββββββββββββββββββββββββββββββββββββ| 5.35M/5.35M [00:00<00:00, 94.6MB/s] Ultralytics 8.3.71 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB) engine/trainer: task=detect, mode=train, model=yolo11n.pt, data=/kaggle/working/smokeandfire-2/data.yaml, epochs=120, time=None, patience=100, batch=32, imgsz=512, save=True, save_period=-1, cache=False, device=None, workers=8, project=None, name=train, exist_ok=False, pretrained=True, optimizer=auto, verbose=True, seed=0, deterministic=True, single_cls=False, rect=False, cos_lr=False, close_mosaic=10, resume=False, amp=True, fraction=1.0, profile=False, freeze=None, multi_scale=False, overlap_mask=True, mask_ratio=4, dropout=0.0, val=True, split=val, save_json=False, save_hybrid=False, conf=None, iou=0.7, max_det=300, half=False, dnn=False, plots=True, source=None, vid_stride=1, stream_buffer=False, visualize=False, augment=True, agnostic_nms=False, classes=None, retina_masks=False, embed=None, show=False, save_frames=False, save_txt=False, save_conf=False, save_crop=False, show_labels=True, show_conf=True, show_boxes=True, line_width=None, format=torchscript, keras=False, optimize=False, int8=False, dynamic=False, simplify=True, opset=None, workspace=None, nms=False, lr0=0.01, lrf=0.01, momentum=0.937, weight_decay=0.0005, warmup_epochs=3.0, warmup_momentum=0.8, warmup_bias_lr=0.1, box=7.5, cls=0.5, dfl=1.5, pose=12.0, kobj=1.0, nbs=64, hsv_h=0.015, hsv_s=0.7, hsv_v=0.4, degrees=0.0, translate=0.1, scale=0.5, shear=0.0, perspective=0.0, flipud=0.0, fliplr=0.5, bgr=0.0, mosaic=1.0, mixup=0.0, copy_paste=0.0, copy_paste_mode=flip, auto_augment=randaugment, erasing=0.4, crop_fraction=1.0, cfg=None, tracker=botsort.yaml, save_dir=runs/detect/train Downloading https://ultralytics.com/assets/Arial.ttf to '/root/.config/Ultralytics/Arial.ttf'... 100%|ββββββββββββββββββββββββββββββββββββββββ| 755k/755k [00:00<00:00, 22.0MB/s] Overriding model.yaml nc=80 with nc=2 from n params module arguments 0 -1 1 464 ultralytics.nn.modules.conv.Conv [3, 16, 3, 2] 1 -1 1 4672 ultralytics.nn.modules.conv.Conv [16, 32, 3, 2] 2 -1 1 6640 ultralytics.nn.modules.block.C3k2 [32, 64, 1, False, 0.25] 3 -1 1 36992 ultralytics.nn.modules.conv.Conv [64, 64, 3, 2] 4 -1 1 26080 ultralytics.nn.modules.block.C3k2 [64, 128, 1, False, 0.25] 5 -1 1 147712 ultralytics.nn.modules.conv.Conv [128, 128, 3, 2] 6 -1 1 87040 ultralytics.nn.modules.block.C3k2 [128, 128, 1, True] 7 -1 1 295424 ultralytics.nn.modules.conv.Conv [128, 256, 3, 2] 8 -1 1 346112 ultralytics.nn.modules.block.C3k2 [256, 256, 1, True] 9 -1 1 164608 ultralytics.nn.modules.block.SPPF [256, 256, 5] 10 -1 1 249728 ultralytics.nn.modules.block.C2PSA [256, 256, 1] 11 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 12 [-1, 6] 1 0 ultralytics.nn.modules.conv.Concat [1] 13 -1 1 111296 ultralytics.nn.modules.block.C3k2 [384, 128, 1, False] 14 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 15 [-1, 4] 1 0 ultralytics.nn.modules.conv.Concat [1] 16 -1 1 32096 ultralytics.nn.modules.block.C3k2 [256, 64, 1, False] 17 -1 1 36992 ultralytics.nn.modules.conv.Conv [64, 64, 3, 2] 18 [-1, 13] 1 0 ultralytics.nn.modules.conv.Concat [1] 19 -1 1 86720 ultralytics.nn.modules.block.C3k2 [192, 128, 1, False] 20 -1 1 147712 ultralytics.nn.modules.conv.Conv [128, 128, 3, 2] 21 [-1, 10] 1 0 ultralytics.nn.modules.conv.Concat [1] 22 -1 1 378880 ultralytics.nn.modules.block.C3k2 [384, 256, 1, True] 23 [16, 19, 22] 1 431062 ultralytics.nn.modules.head.Detect [2, [64, 128, 256]] YOLO11n summary: 319 layers, 2,590,230 parameters, 2,590,214 gradients, 6.4 GFLOPs Transferred 448/499 items from pretrained weights TensorBoard: Start with 'tensorboard --logdir runs/detect/train', view at http://localhost:6006/ Freezing layer 'model.23.dfl.conv.weight' AMP: running Automatic Mixed Precision (AMP) checks... AMP: checks passed β train: Scanning /kaggle/working/smokeandfire-2/train/labels... 8243 images, 11 b train: New cache created: /kaggle/working/smokeandfire-2/train/labels.cache WARNING β οΈ Box and segment counts should be equal, but got len(segments) = 12, len(boxes) = 16365. To resolve this only boxes will be used and all segments will be removed. To avoid this please supply either a detect or segment dataset, not a detect-segment mixed dataset. /opt/conda/lib/python3.10/site-packages/albumentations/__init__.py:24: UserWarning: A new version of Albumentations is available: 2.0.3 (you have 1.4.21). Upgrade using: pip install -U albumentations. To disable automatic update checks, set the environment variable NO_ALBUMENTATIONS_UPDATE to 1. check_for_updates() albumentations: Blur(p=0.01, blur_limit=(3, 7)), MedianBlur(p=0.01, blur_limit=(3, 7)), ToGray(p=0.01, num_output_channels=3, method='weighted_average'), CLAHE(p=0.01, clip_limit=(1.0, 4.0), tile_grid_size=(8, 8)) /opt/conda/lib/python3.10/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock. self.pid = os.fork() val: Scanning /kaggle/working/smokeandfire-2/valid/labels... 1062 images, 3 back val: New cache created: /kaggle/working/smokeandfire-2/valid/labels.cache /opt/conda/lib/python3.10/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock. self.pid = os.fork() Plotting labels to runs/detect/train/labels.jpg... optimizer: 'optimizer=auto' found, ignoring 'lr0=0.01' and 'momentum=0.937' and determining best 'optimizer', 'lr0' and 'momentum' automatically... optimizer: SGD(lr=0.01, momentum=0.9) with parameter groups 81 weight(decay=0.0), 88 weight(decay=0.0005), 87 bias(decay=0.0) TensorBoard: model graph visualization added β Image sizes 512 train, 512 val Using 4 dataloader workers Logging results to runs/detect/train Starting training for 120 epochs... Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 1/120 3.26G 1.436 2.61 1.466 110 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.707 0.572 0.64 0.44 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 2/120 3.31G 1.362 1.72 1.379 68 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.814 0.696 0.749 0.537 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 3/120 3.31G 1.419 1.631 1.417 42 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.667 0.405 0.446 0.296 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 4/120 3.3G 1.478 1.639 1.473 66 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.768 0.611 0.669 0.478 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 5/120 3.3G 1.433 1.515 1.448 69 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.773 0.678 0.712 0.518 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 6/120 3.3G 1.398 1.441 1.422 65 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.817 0.735 0.767 0.556 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 7/120 3.29G 1.358 1.372 1.403 66 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.824 0.697 0.742 0.548 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 8/120 3.3G 1.339 1.326 1.384 54 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.82 0.766 0.789 0.586 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 9/120 3.33G 1.305 1.278 1.371 49 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.848 0.774 0.8 0.595 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 10/120 3.29G 1.305 1.286 1.371 79 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.834 0.789 0.807 0.605 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 11/120 3.27G 1.265 1.225 1.346 59 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.858 0.809 0.819 0.611 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 12/120 3.28G 1.259 1.214 1.339 81 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.859 0.783 0.809 0.615 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 13/120 3.3G 1.257 1.196 1.343 57 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.839 0.778 0.804 0.614 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 14/120 3.31G 1.238 1.175 1.326 66 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.87 0.791 0.826 0.623 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 15/120 3.27G 1.227 1.151 1.325 78 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.857 0.815 0.83 0.635 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 16/120 3.27G 1.213 1.123 1.311 68 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.854 0.812 0.843 0.637 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 17/120 3.28G 1.223 1.133 1.314 68 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.865 0.824 0.844 0.641 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 18/120 3.29G 1.187 1.096 1.297 75 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.87 0.825 0.834 0.639 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 19/120 3.29G 1.207 1.109 1.304 53 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.871 0.841 0.857 0.658 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 20/120 3.3G 1.18 1.08 1.292 82 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.879 0.831 0.851 0.663 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 21/120 3.27G 1.187 1.085 1.298 62 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.871 0.831 0.85 0.658 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 22/120 3.27G 1.19 1.068 1.288 79 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.882 0.832 0.848 0.659 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 23/120 3.27G 1.165 1.049 1.281 71 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.869 0.839 0.843 0.659 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 24/120 3.3G 1.162 1.047 1.276 75 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.884 0.822 0.852 0.66 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 25/120 3.3G 1.148 1.021 1.27 51 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.884 0.833 0.855 0.664 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 26/120 3.3G 1.156 1.037 1.272 81 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.882 0.838 0.852 0.666 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 27/120 3.3G 1.146 1.022 1.267 73 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.879 0.846 0.857 0.67 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 28/120 3.3G 1.136 1.008 1.261 58 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.872 0.838 0.855 0.671 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 29/120 3.31G 1.137 1.009 1.264 68 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.88 0.839 0.855 0.676 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 30/120 3.27G 1.142 1.005 1.261 40 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.872 0.845 0.854 0.668 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 31/120 3.3G 1.128 0.9905 1.255 47 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.898 0.846 0.863 0.684 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 32/120 3.3G 1.113 0.9722 1.245 68 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.884 0.853 0.863 0.675 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 33/120 3.3G 1.12 0.9904 1.249 77 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.879 0.856 0.864 0.679 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 34/120 3.3G 1.099 0.9649 1.235 47 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.893 0.847 0.857 0.679 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 35/120 3.28G 1.111 0.9657 1.241 54 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.899 0.844 0.858 0.685 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 36/120 3.3G 1.118 0.9633 1.244 70 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.896 0.841 0.858 0.679 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 37/120 3.27G 1.111 0.9528 1.238 41 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.897 0.852 0.868 0.686 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 38/120 3.25G 1.089 0.9448 1.23 62 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.898 0.862 0.867 0.69 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 39/120 3.28G 1.09 0.9326 1.227 56 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.89 0.856 0.871 0.688 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 40/120 3.27G 1.088 0.9277 1.225 77 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.892 0.851 0.867 0.691 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 41/120 3.31G 1.085 0.9331 1.223 54 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.896 0.855 0.862 0.685 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 42/120 3.3G 1.086 0.921 1.225 61 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.888 0.863 0.868 0.69 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 43/120 3.27G 1.078 0.9189 1.22 56 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.894 0.865 0.864 0.691 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 44/120 3.3G 1.077 0.9252 1.222 65 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.897 0.864 0.872 0.692 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 45/120 3.31G 1.072 0.9073 1.22 61 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.859 0.868 0.691 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 46/120 3.27G 1.066 0.8929 1.21 74 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.899 0.859 0.872 0.693 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 47/120 3.29G 1.06 0.8987 1.211 55 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.906 0.862 0.873 0.694 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 48/120 3.3G 1.082 0.9064 1.219 63 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.905 0.856 0.869 0.693 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 49/120 3.31G 1.059 0.8844 1.205 74 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.901 0.86 0.871 0.696 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 50/120 3.29G 1.06 0.8844 1.207 70 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.903 0.86 0.872 0.694 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 51/120 3.31G 1.041 0.874 1.193 69 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.905 0.866 0.878 0.703 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 52/120 3.31G 1.044 0.8695 1.201 71 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.897 0.862 0.872 0.702 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 53/120 3.29G 1.049 0.8753 1.202 56 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.906 0.866 0.876 0.703 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 54/120 3.31G 1.045 0.8659 1.2 79 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.904 0.862 0.874 0.701 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 55/120 3.31G 1.038 0.8652 1.195 59 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.905 0.866 0.871 0.701 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 56/120 3.27G 1.031 0.8554 1.193 76 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.903 0.868 0.873 0.698 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 57/120 3.31G 1.037 0.8555 1.191 63 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.867 0.873 0.7 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 58/120 3.27G 1.034 0.8462 1.191 98 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.901 0.859 0.872 0.703 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 59/120 3.29G 1.032 0.8382 1.19 120 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.901 0.869 0.878 0.704 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 60/120 3.3G 1.027 0.8497 1.188 73 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.901 0.868 0.877 0.703 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 61/120 3.3G 1.029 0.8432 1.185 63 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.905 0.863 0.878 0.708 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 62/120 3.3G 1.016 0.827 1.179 63 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.904 0.869 0.877 0.702 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 63/120 3.3G 1.016 0.8355 1.184 73 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.909 0.864 0.879 0.707 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 64/120 3.29G 1.004 0.8192 1.171 61 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.896 0.868 0.878 0.707 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 65/120 3.27G 1.02 0.8247 1.176 97 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.899 0.869 0.875 0.705 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 66/120 3.27G 1.018 0.8184 1.182 56 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.903 0.872 0.876 0.706 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 67/120 3.29G 0.9944 0.7994 1.167 79 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.902 0.867 0.873 0.704 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 68/120 3.3G 1.006 0.802 1.174 78 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.908 0.864 0.874 0.708 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 69/120 3.3G 0.9867 0.788 1.159 68 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.904 0.865 0.874 0.705 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 70/120 3.29G 0.9927 0.8019 1.162 75 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.898 0.867 0.874 0.707 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 71/120 3.27G 0.9946 0.7969 1.161 79 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.906 0.865 0.875 0.708 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 72/120 3.3G 0.9899 0.7875 1.161 64 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.906 0.867 0.875 0.71 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 73/120 3.27G 0.9975 0.7977 1.169 44 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.906 0.866 0.875 0.71 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 74/120 3.29G 0.9864 0.7849 1.16 72 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.904 0.868 0.876 0.713 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 75/120 3.3G 0.9777 0.7802 1.158 69 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.905 0.868 0.878 0.713 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 76/120 3.27G 0.9854 0.7775 1.157 63 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.907 0.869 0.878 0.713 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 77/120 3.3G 0.9822 0.7828 1.159 65 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.906 0.869 0.878 0.714 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 78/120 3.27G 0.9772 0.7644 1.151 68 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.907 0.869 0.88 0.716 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 79/120 3.27G 0.9696 0.764 1.148 54 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.907 0.866 0.878 0.714 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 80/120 3.29G 0.9681 0.7653 1.148 64 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.909 0.868 0.879 0.716 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 81/120 3.31G 0.9664 0.7535 1.142 84 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.906 0.87 0.878 0.715 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 82/120 3.31G 0.9643 0.754 1.145 85 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.908 0.871 0.877 0.715 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 83/120 3.29G 0.9526 0.7361 1.138 68 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.905 0.872 0.877 0.714 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 84/120 3.27G 0.9588 0.7382 1.138 77 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.907 0.871 0.877 0.715 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 85/120 3.29G 0.9594 0.7385 1.143 82 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.91 0.873 0.88 0.718 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 86/120 3.3G 0.958 0.7438 1.141 62 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.907 0.873 0.88 0.718 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 87/120 3.28G 0.9399 0.7295 1.131 86 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.904 0.869 0.879 0.717 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 88/120 3.31G 0.9461 0.7319 1.137 50 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.905 0.867 0.88 0.718 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 89/120 3.3G 0.9431 0.7254 1.129 72 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.907 0.868 0.881 0.719 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 90/120 3.27G 0.9373 0.7128 1.125 70 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.907 0.87 0.88 0.718 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 91/120 3.29G 0.9316 0.7105 1.125 46 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.908 0.868 0.88 0.718 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 92/120 3.3G 0.9318 0.7071 1.13 60 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.906 0.867 0.88 0.718 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 93/120 3.31G 0.9307 0.7139 1.125 61 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.91 0.867 0.882 0.719 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 94/120 3.29G 0.9355 0.7159 1.127 81 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.908 0.87 0.881 0.72 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 95/120 3.27G 0.9325 0.7044 1.123 92 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.906 0.871 0.883 0.721 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 96/120 3.29G 0.9364 0.7062 1.124 75 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.906 0.869 0.881 0.721 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 97/120 3.3G 0.9243 0.6948 1.117 55 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.906 0.867 0.881 0.72 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 98/120 3.3G 0.9179 0.6887 1.119 59 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.909 0.868 0.882 0.722 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 99/120 3.31G 0.9128 0.6816 1.113 65 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.909 0.868 0.881 0.72 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 100/120 3.25G 0.9169 0.6878 1.114 74 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.91 0.865 0.88 0.719 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 101/120 3.27G 0.9099 0.6768 1.111 40 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.909 0.866 0.878 0.719 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 102/120 3.34G 0.8971 0.6639 1.102 76 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.907 0.867 0.877 0.717 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 103/120 3.31G 0.9041 0.6708 1.109 65 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.908 0.867 0.878 0.718 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 104/120 3.3G 0.8913 0.661 1.101 67 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.907 0.866 0.878 0.717 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 105/120 3.28G 0.8959 0.6578 1.102 71 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.909 0.867 0.878 0.717 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 106/120 3.3G 0.89 0.6575 1.102 71 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.907 0.867 0.879 0.717 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 107/120 3.31G 0.905 0.664 1.106 61 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.908 0.865 0.878 0.717 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 108/120 3.3G 0.8902 0.6515 1.1 77 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.908 0.865 0.877 0.717 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 109/120 3.31G 0.8954 0.6533 1.101 73 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.907 0.864 0.878 0.717 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 110/120 3.29G 0.8811 0.646 1.097 60 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.908 0.867 0.878 0.717 Closing dataloader mosaic albumentations: Blur(p=0.01, blur_limit=(3, 7)), MedianBlur(p=0.01, blur_limit=(3, 7)), ToGray(p=0.01, num_output_channels=3, method='weighted_average'), CLAHE(p=0.01, clip_limit=(1.0, 4.0), tile_grid_size=(8, 8)) /opt/conda/lib/python3.10/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock. self.pid = os.fork() Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 111/120 3.27G 0.8847 0.5868 1.083 45 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.907 0.866 0.878 0.717 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 112/120 3.25G 0.8675 0.5623 1.075 28 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.905 0.866 0.877 0.717 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 113/120 3.25G 0.8625 0.5502 1.07 34 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.904 0.866 0.878 0.717 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 114/120 3.27G 0.8532 0.5476 1.065 34 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.903 0.866 0.877 0.717 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 115/120 3.27G 0.8476 0.537 1.063 41 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.902 0.866 0.878 0.717 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 116/120 3.27G 0.8397 0.5279 1.059 33 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.903 0.866 0.878 0.717 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 117/120 3.27G 0.8393 0.5274 1.059 34 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.903 0.865 0.878 0.717 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 118/120 3.27G 0.8387 0.5298 1.06 43 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.903 0.864 0.878 0.717 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 119/120 3.3G 0.8296 0.5176 1.055 39 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.904 0.864 0.878 0.717 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 120/120 3.28G 0.824 0.5132 1.054 35 512: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.905 0.865 0.878 0.716 120 epochs completed in 2.325 hours. Optimizer stripped from runs/detect/train/weights/last.pt, 5.5MB Optimizer stripped from runs/detect/train/weights/best.pt, 5.5MB Validating runs/detect/train/weights/best.pt... Ultralytics 8.3.71 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB) YOLO11n summary (fused): 238 layers, 2,582,542 parameters, 0 gradients, 6.3 GFLOPs Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.871 0.876 0.715 Fire 859 1320 0.917 0.943 0.94 0.773 Smoke 559 761 0.874 0.8 0.812 0.657 Speed: 0.1ms preprocess, 3.5ms inference, 0.0ms loss, 0.8ms postprocess per image Results saved to runs/detect/train π‘ Learn more at https://docs.ultralytics.com/modes/train
yolo v 12ΒΆ
InΒ [Β ]:
!pip install roboflow
from roboflow import Roboflow
rf = Roboflow(api_key="GZBlBPlRdRvB6uR7BqaS")
project = rf.workspace("detection-e83li").project("smokeandfire")
version = project.version(2)
dataset = version.download("yolov12")
InΒ [3]:
!yolo train model=yolo12s.pt data='/kaggle/working/smokeandfire-2/data.yaml' epochs=20 imgsz=640 batch=32 augment=True
Creating new Ultralytics Settings v0.0.6 file β View Ultralytics Settings with 'yolo settings' or at '/root/.config/Ultralytics/settings.json' Update Settings with 'yolo settings key=value', i.e. 'yolo settings runs_dir=path/to/dir'. For help see https://docs.ultralytics.com/quickstart/#ultralytics-settings. Downloading https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo12s.pt to 'yolo12s.pt'... 100%|ββββββββββββββββββββββββββββββββββββββ| 18.1M/18.1M [00:01<00:00, 18.0MB/s] Ultralytics 8.3.78 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB) engine/trainer: task=detect, mode=train, model=yolo12s.pt, data=/kaggle/working/smokeandfire-2/data.yaml, epochs=20, time=None, patience=100, batch=32, imgsz=640, save=True, save_period=-1, cache=False, device=None, workers=8, project=None, name=train, exist_ok=False, pretrained=True, optimizer=auto, verbose=True, seed=0, deterministic=True, single_cls=False, rect=False, cos_lr=False, close_mosaic=10, resume=False, amp=True, fraction=1.0, profile=False, freeze=None, multi_scale=False, overlap_mask=True, mask_ratio=4, dropout=0.0, val=True, split=val, save_json=False, save_hybrid=False, conf=None, iou=0.7, max_det=300, half=False, dnn=False, plots=True, source=None, vid_stride=1, stream_buffer=False, visualize=False, augment=True, agnostic_nms=False, classes=None, retina_masks=False, embed=None, show=False, save_frames=False, save_txt=False, save_conf=False, save_crop=False, show_labels=True, show_conf=True, show_boxes=True, line_width=None, format=torchscript, keras=False, optimize=False, int8=False, dynamic=False, simplify=True, opset=None, workspace=None, nms=False, lr0=0.01, lrf=0.01, momentum=0.937, weight_decay=0.0005, warmup_epochs=3.0, warmup_momentum=0.8, warmup_bias_lr=0.1, box=7.5, cls=0.5, dfl=1.5, pose=12.0, kobj=1.0, nbs=64, hsv_h=0.015, hsv_s=0.7, hsv_v=0.4, degrees=0.0, translate=0.1, scale=0.5, shear=0.0, perspective=0.0, flipud=0.0, fliplr=0.5, bgr=0.0, mosaic=1.0, mixup=0.0, copy_paste=0.0, copy_paste_mode=flip, auto_augment=randaugment, erasing=0.4, crop_fraction=1.0, cfg=None, tracker=botsort.yaml, save_dir=runs/detect/train Downloading https://ultralytics.com/assets/Arial.ttf to '/root/.config/Ultralytics/Arial.ttf'... 100%|βββββββββββββββββββββββββββββββββββββββββ| 755k/755k [00:00<00:00, 142MB/s] Overriding model.yaml nc=80 with nc=2 from n params module arguments 0 -1 1 928 ultralytics.nn.modules.conv.Conv [3, 32, 3, 2] 1 -1 1 18560 ultralytics.nn.modules.conv.Conv [32, 64, 3, 2] 2 -1 1 26080 ultralytics.nn.modules.block.C3k2 [64, 128, 1, False, 0.25] 3 -1 1 147712 ultralytics.nn.modules.conv.Conv [128, 128, 3, 2] 4 -1 1 103360 ultralytics.nn.modules.block.C3k2 [128, 256, 1, False, 0.25] 5 -1 1 590336 ultralytics.nn.modules.conv.Conv [256, 256, 3, 2] 6 -1 2 689408 ultralytics.nn.modules.block.A2C2f [256, 256, 2, True, 4] 7 -1 1 1180672 ultralytics.nn.modules.conv.Conv [256, 512, 3, 2] 8 -1 2 2689536 ultralytics.nn.modules.block.A2C2f [512, 512, 2, True, 1] 9 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 10 [-1, 6] 1 0 ultralytics.nn.modules.conv.Concat [1] 11 -1 1 345856 ultralytics.nn.modules.block.A2C2f [768, 256, 1, False, -1] 12 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 13 [-1, 4] 1 0 ultralytics.nn.modules.conv.Concat [1] 14 -1 1 95104 ultralytics.nn.modules.block.A2C2f [512, 128, 1, False, -1] 15 -1 1 147712 ultralytics.nn.modules.conv.Conv [128, 128, 3, 2] 16 [-1, 11] 1 0 ultralytics.nn.modules.conv.Concat [1] 17 -1 1 296704 ultralytics.nn.modules.block.A2C2f [384, 256, 1, False, -1] 18 -1 1 590336 ultralytics.nn.modules.conv.Conv [256, 256, 3, 2] 19 [-1, 8] 1 0 ultralytics.nn.modules.conv.Concat [1] 20 -1 1 1511424 ultralytics.nn.modules.block.C3k2 [768, 512, 1, True] 21 [14, 17, 20] 1 820182 ultralytics.nn.modules.head.Detect [2, [128, 256, 512]] YOLOv12s summary: 272 layers, 9,253,910 parameters, 9,253,894 gradients, 21.5 GFLOPs Transferred 685/691 items from pretrained weights TensorBoard: Start with 'tensorboard --logdir runs/detect/train', view at http://localhost:6006/ Freezing layer 'model.21.dfl.conv.weight' AMP: running Automatic Mixed Precision (AMP) checks... Downloading https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11n.pt to 'yolo11n.pt'... 100%|βββββββββββββββββββββββββββββββββββββββ| 5.35M/5.35M [00:00<00:00, 269MB/s] AMP: checks passed β train: Scanning /kaggle/working/smokeandfire-2/train/labels... 8243 images, 11 b train: New cache created: /kaggle/working/smokeandfire-2/train/labels.cache WARNING β οΈ Box and segment counts should be equal, but got len(segments) = 12, len(boxes) = 16365. To resolve this only boxes will be used and all segments will be removed. To avoid this please supply either a detect or segment dataset, not a detect-segment mixed dataset. /opt/conda/lib/python3.10/site-packages/albumentations/__init__.py:24: UserWarning: A new version of Albumentations is available: 2.0.4 (you have 1.4.21). Upgrade using: pip install -U albumentations. To disable automatic update checks, set the environment variable NO_ALBUMENTATIONS_UPDATE to 1. check_for_updates() albumentations: Blur(p=0.01, blur_limit=(3, 7)), MedianBlur(p=0.01, blur_limit=(3, 7)), ToGray(p=0.01, num_output_channels=3, method='weighted_average'), CLAHE(p=0.01, clip_limit=(1.0, 4.0), tile_grid_size=(8, 8)) /opt/conda/lib/python3.10/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock. self.pid = os.fork() val: Scanning /kaggle/working/smokeandfire-2/valid/labels... 1062 images, 3 back val: New cache created: /kaggle/working/smokeandfire-2/valid/labels.cache /opt/conda/lib/python3.10/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock. self.pid = os.fork() Plotting labels to runs/detect/train/labels.jpg... optimizer: 'optimizer=auto' found, ignoring 'lr0=0.01' and 'momentum=0.937' and determining best 'optimizer', 'lr0' and 'momentum' automatically... optimizer: AdamW(lr=0.001667, momentum=0.9) with parameter groups 113 weight(decay=0.0), 120 weight(decay=0.0005), 119 bias(decay=0.0) TensorBoard: model graph visualization added β Image sizes 640 train, 640 val Using 4 dataloader workers Logging results to runs/detect/train Starting training for 20 epochs... Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 1/20 12.4G 1.512 2.271 1.654 111 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.26 0.492 0.251 0.148 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 2/20 12G 1.532 1.74 1.67 68 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.791 0.667 0.72 0.508 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 3/20 12.3G 1.471 1.595 1.618 42 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.721 0.664 0.71 0.491 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 4/20 12G 1.417 1.517 1.582 67 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.831 0.648 0.72 0.529 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 5/20 12G 1.362 1.416 1.538 69 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.809 0.744 0.781 0.57 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 6/20 12G 1.329 1.358 1.515 65 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.849 0.796 0.814 0.607 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 7/20 12G 1.273 1.28 1.477 66 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.863 0.786 0.817 0.613 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 8/20 12G 1.252 1.241 1.462 54 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.865 0.799 0.826 0.623 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 9/20 12.1G 1.221 1.19 1.443 49 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.856 0.79 0.812 0.625 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 10/20 12.2G 1.226 1.187 1.452 79 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.828 0.795 0.809 0.616 Closing dataloader mosaic albumentations: Blur(p=0.01, blur_limit=(3, 7)), MedianBlur(p=0.01, blur_limit=(3, 7)), ToGray(p=0.01, num_output_channels=3, method='weighted_average'), CLAHE(p=0.01, clip_limit=(1.0, 4.0), tile_grid_size=(8, 8)) /opt/conda/lib/python3.10/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock. self.pid = os.fork() Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 11/20 12G 1.261 1.134 1.493 41 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.862 0.814 0.831 0.641 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 12/20 12G 1.241 1.088 1.477 38 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.873 0.827 0.843 0.64 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 13/20 12G 1.224 1.063 1.461 34 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.878 0.834 0.845 0.666 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 14/20 12G 1.193 1.029 1.436 39 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.877 0.85 0.856 0.662 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 15/20 12G 1.165 0.9949 1.416 25 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.882 0.852 0.861 0.676 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 16/20 12G 1.152 0.955 1.406 35 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.866 0.849 0.85 0.667 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 17/20 12G 1.117 0.919 1.385 45 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.876 0.854 0.858 0.684 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 18/20 12G 1.095 0.8873 1.367 33 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.898 0.859 0.865 0.693 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 19/20 12G 1.076 0.8529 1.352 31 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.898 0.86 0.871 0.699 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 20/20 12G 1.048 0.8218 1.335 37 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.858 0.87 0.703 20 epochs completed in 1.573 hours. Optimizer stripped from runs/detect/train/weights/last.pt, 18.9MB Optimizer stripped from runs/detect/train/weights/best.pt, 18.9MB Validating runs/detect/train/weights/best.pt... Ultralytics 8.3.78 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB) YOLOv12s summary (fused): 159 layers, 9,231,654 parameters, 0 gradients, 21.2 GFLOPs Class Images Instances Box(P R mAP50 m all 1062 2081 0.878 0.858 0.866 0.696 Fire 859 1320 0.899 0.936 0.933 0.754 Smoke 559 761 0.856 0.779 0.8 0.639 Speed: 0.2ms preprocess, 15.5ms inference, 0.0ms loss, 0.8ms postprocess per image Results saved to runs/detect/train π‘ Learn more at https://docs.ultralytics.com/modes/train
InΒ [3]:
!yolo train model=yolo12s.pt data='/kaggle/working/smokeandfire-2/data.yaml' epochs=80 imgsz=640 batch=32 augment=True lr0=0.005 lrf=0.01 weight_decay=0.0005 optimizer=SGD mosaic=True hsv_h=0.015 hsv_s=0.7 hsv_v=0.4 degrees=15 translate=0.1 scale=0.5 shear=0.2 flipud=0.5 fliplr=0.5 mixup=0.2
Creating new Ultralytics Settings v0.0.6 file β View Ultralytics Settings with 'yolo settings' or at '/root/.config/Ultralytics/settings.json' Update Settings with 'yolo settings key=value', i.e. 'yolo settings runs_dir=path/to/dir'. For help see https://docs.ultralytics.com/quickstart/#ultralytics-settings. Downloading https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo12s.pt to 'yolo12s.pt'... 100%|βββββββββββββββββββββββββββββββββββββββ| 18.1M/18.1M [00:00<00:00, 174MB/s] Ultralytics 8.3.78 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB) engine/trainer: task=detect, mode=train, model=yolo12s.pt, data=/kaggle/working/smokeandfire-2/data.yaml, epochs=80, time=None, patience=100, batch=32, imgsz=640, save=True, save_period=-1, cache=False, device=None, workers=8, project=None, name=train, exist_ok=False, pretrained=True, optimizer=SGD, verbose=True, seed=0, deterministic=True, single_cls=False, rect=False, cos_lr=False, close_mosaic=10, resume=False, amp=True, fraction=1.0, profile=False, freeze=None, multi_scale=False, overlap_mask=True, mask_ratio=4, dropout=0.0, val=True, split=val, save_json=False, save_hybrid=False, conf=None, iou=0.7, max_det=300, half=False, dnn=False, plots=True, source=None, vid_stride=1, stream_buffer=False, visualize=False, augment=True, agnostic_nms=False, classes=None, retina_masks=False, embed=None, show=False, save_frames=False, save_txt=False, save_conf=False, save_crop=False, show_labels=True, show_conf=True, show_boxes=True, line_width=None, format=torchscript, keras=False, optimize=False, int8=False, dynamic=False, simplify=True, opset=None, workspace=None, nms=False, lr0=0.005, lrf=0.01, momentum=0.937, weight_decay=0.0005, warmup_epochs=3.0, warmup_momentum=0.8, warmup_bias_lr=0.1, box=7.5, cls=0.5, dfl=1.5, pose=12.0, kobj=1.0, nbs=64, hsv_h=0.015, hsv_s=0.7, hsv_v=0.4, degrees=15, translate=0.1, scale=0.5, shear=0.2, perspective=0.0, flipud=0.5, fliplr=0.5, bgr=0.0, mosaic=True, mixup=0.2, copy_paste=0.0, copy_paste_mode=flip, auto_augment=randaugment, erasing=0.4, crop_fraction=1.0, cfg=None, tracker=botsort.yaml, save_dir=runs/detect/train Downloading https://ultralytics.com/assets/Arial.ttf to '/root/.config/Ultralytics/Arial.ttf'... 100%|ββββββββββββββββββββββββββββββββββββββββ| 755k/755k [00:00<00:00, 16.9MB/s] Overriding model.yaml nc=80 with nc=2 from n params module arguments 0 -1 1 928 ultralytics.nn.modules.conv.Conv [3, 32, 3, 2] 1 -1 1 18560 ultralytics.nn.modules.conv.Conv [32, 64, 3, 2] 2 -1 1 26080 ultralytics.nn.modules.block.C3k2 [64, 128, 1, False, 0.25] 3 -1 1 147712 ultralytics.nn.modules.conv.Conv [128, 128, 3, 2] 4 -1 1 103360 ultralytics.nn.modules.block.C3k2 [128, 256, 1, False, 0.25] 5 -1 1 590336 ultralytics.nn.modules.conv.Conv [256, 256, 3, 2] 6 -1 2 689408 ultralytics.nn.modules.block.A2C2f [256, 256, 2, True, 4] 7 -1 1 1180672 ultralytics.nn.modules.conv.Conv [256, 512, 3, 2] 8 -1 2 2689536 ultralytics.nn.modules.block.A2C2f [512, 512, 2, True, 1] 9 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 10 [-1, 6] 1 0 ultralytics.nn.modules.conv.Concat [1] 11 -1 1 345856 ultralytics.nn.modules.block.A2C2f [768, 256, 1, False, -1] 12 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 13 [-1, 4] 1 0 ultralytics.nn.modules.conv.Concat [1] 14 -1 1 95104 ultralytics.nn.modules.block.A2C2f [512, 128, 1, False, -1] 15 -1 1 147712 ultralytics.nn.modules.conv.Conv [128, 128, 3, 2] 16 [-1, 11] 1 0 ultralytics.nn.modules.conv.Concat [1] 17 -1 1 296704 ultralytics.nn.modules.block.A2C2f [384, 256, 1, False, -1] 18 -1 1 590336 ultralytics.nn.modules.conv.Conv [256, 256, 3, 2] 19 [-1, 8] 1 0 ultralytics.nn.modules.conv.Concat [1] 20 -1 1 1511424 ultralytics.nn.modules.block.C3k2 [768, 512, 1, True] 21 [14, 17, 20] 1 820182 ultralytics.nn.modules.head.Detect [2, [128, 256, 512]] YOLOv12s summary: 272 layers, 9,253,910 parameters, 9,253,894 gradients, 21.5 GFLOPs Transferred 685/691 items from pretrained weights TensorBoard: Start with 'tensorboard --logdir runs/detect/train', view at http://localhost:6006/ Freezing layer 'model.21.dfl.conv.weight' AMP: running Automatic Mixed Precision (AMP) checks... Downloading https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11n.pt to 'yolo11n.pt'... 100%|ββββββββββββββββββββββββββββββββββββββ| 5.35M/5.35M [00:00<00:00, 72.9MB/s] AMP: checks passed β train: Scanning /kaggle/working/smokeandfire-2/train/labels... 8243 images, 11 b train: New cache created: /kaggle/working/smokeandfire-2/train/labels.cache WARNING β οΈ Box and segment counts should be equal, but got len(segments) = 12, len(boxes) = 16365. To resolve this only boxes will be used and all segments will be removed. To avoid this please supply either a detect or segment dataset, not a detect-segment mixed dataset. /opt/conda/lib/python3.10/site-packages/albumentations/__init__.py:24: UserWarning: A new version of Albumentations is available: 2.0.4 (you have 1.4.21). Upgrade using: pip install -U albumentations. To disable automatic update checks, set the environment variable NO_ALBUMENTATIONS_UPDATE to 1. check_for_updates() albumentations: Blur(p=0.01, blur_limit=(3, 7)), MedianBlur(p=0.01, blur_limit=(3, 7)), ToGray(p=0.01, num_output_channels=3, method='weighted_average'), CLAHE(p=0.01, clip_limit=(1.0, 4.0), tile_grid_size=(8, 8)) /opt/conda/lib/python3.10/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock. self.pid = os.fork() val: Scanning /kaggle/working/smokeandfire-2/valid/labels... 1062 images, 3 back val: New cache created: /kaggle/working/smokeandfire-2/valid/labels.cache /opt/conda/lib/python3.10/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock. self.pid = os.fork() Plotting labels to runs/detect/train/labels.jpg... optimizer: SGD(lr=0.005, momentum=0.937) with parameter groups 113 weight(decay=0.0), 120 weight(decay=0.0005), 119 bias(decay=0.0) TensorBoard: model graph visualization added β Image sizes 640 train, 640 val Using 4 dataloader workers Logging results to runs/detect/train Starting training for 80 epochs... Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 1/80 12.4G 1.606 2.359 1.799 96 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.823 0.738 0.77 0.541 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 2/80 12.2G 1.42 1.452 1.603 76 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.846 0.759 0.798 0.551 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 3/80 12.2G 1.473 1.497 1.645 53 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.814 0.686 0.721 0.497 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 4/80 12G 1.532 1.618 1.704 90 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.827 0.675 0.716 0.501 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 5/80 12G 1.5 1.554 1.68 95 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.834 0.753 0.794 0.56 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 6/80 12G 1.462 1.485 1.659 62 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.839 0.784 0.793 0.573 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 7/80 12G 1.43 1.452 1.63 72 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.845 0.753 0.789 0.562 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 8/80 12G 1.417 1.417 1.623 82 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.824 0.777 0.814 0.595 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 9/80 12G 1.386 1.388 1.601 54 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.852 0.804 0.823 0.606 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 10/80 12G 1.372 1.352 1.595 91 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.849 0.806 0.822 0.617 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 11/80 12G 1.351 1.324 1.57 88 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.857 0.817 0.831 0.621 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 12/80 12G 1.343 1.305 1.564 112 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.869 0.808 0.832 0.612 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 13/80 12.1G 1.338 1.293 1.558 76 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.854 0.81 0.828 0.613 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 14/80 12G 1.318 1.272 1.542 80 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.858 0.816 0.832 0.625 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 15/80 12G 1.314 1.268 1.547 114 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.868 0.837 0.845 0.637 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 16/80 12G 1.294 1.25 1.534 91 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.853 0.819 0.828 0.627 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 17/80 12G 1.285 1.233 1.528 65 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.867 0.827 0.835 0.632 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 18/80 12G 1.285 1.222 1.528 137 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.873 0.814 0.837 0.632 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 19/80 12G 1.279 1.211 1.524 92 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.865 0.832 0.841 0.637 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 20/80 12G 1.278 1.201 1.523 92 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.884 0.835 0.847 0.653 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 21/80 12G 1.255 1.163 1.498 75 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.885 0.847 0.852 0.642 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 22/80 12G 1.26 1.183 1.509 87 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.866 0.826 0.841 0.65 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 23/80 12G 1.243 1.159 1.5 117 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.883 0.839 0.856 0.654 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 24/80 12G 1.239 1.152 1.492 129 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.887 0.839 0.856 0.665 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 25/80 12G 1.237 1.124 1.481 80 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.886 0.85 0.864 0.668 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 26/80 12G 1.24 1.136 1.489 78 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.879 0.845 0.862 0.661 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 27/80 12G 1.213 1.114 1.467 75 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.852 0.867 0.675 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 28/80 12G 1.223 1.117 1.475 71 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.898 0.85 0.863 0.67 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 29/80 12G 1.215 1.101 1.472 71 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.893 0.853 0.868 0.672 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 30/80 12G 1.207 1.093 1.46 70 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.891 0.852 0.86 0.675 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 31/80 12G 1.192 1.072 1.451 68 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.887 0.852 0.863 0.671 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 32/80 12G 1.194 1.086 1.454 75 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.892 0.853 0.866 0.675 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 33/80 12G 1.194 1.066 1.451 97 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.849 0.871 0.682 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 34/80 12G 1.192 1.054 1.446 72 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.891 0.857 0.869 0.682 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 35/80 12G 1.176 1.058 1.443 72 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.849 0.866 0.684 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 36/80 12G 1.161 1.036 1.43 100 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.892 0.853 0.867 0.685 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 37/80 12G 1.172 1.033 1.431 81 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.898 0.853 0.87 0.688 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 38/80 12G 1.177 1.053 1.445 74 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.9 0.859 0.872 0.693 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 39/80 12G 1.16 1.017 1.426 98 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.893 0.859 0.867 0.69 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 40/80 12G 1.15 0.999 1.414 77 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.902 0.856 0.872 0.691 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 41/80 12G 1.16 1.006 1.424 78 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.9 0.861 0.866 0.696 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 42/80 12G 1.149 0.999 1.414 111 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.904 0.864 0.875 0.698 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 43/80 12G 1.153 1.005 1.412 111 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.903 0.855 0.872 0.697 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 44/80 12G 1.135 0.9739 1.412 67 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.908 0.86 0.875 0.695 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 45/80 12G 1.138 0.9789 1.412 79 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.898 0.866 0.877 0.697 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 46/80 12G 1.131 0.9579 1.399 94 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.901 0.861 0.874 0.697 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 47/80 12G 1.114 0.9625 1.39 77 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.908 0.861 0.88 0.703 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 48/80 12G 1.134 0.9643 1.402 57 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.902 0.864 0.88 0.706 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 49/80 12G 1.113 0.9482 1.391 96 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.902 0.864 0.877 0.704 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 50/80 12G 1.114 0.9463 1.396 91 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.903 0.86 0.88 0.709 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 51/80 12G 1.113 0.9379 1.387 76 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.906 0.861 0.884 0.708 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 52/80 12G 1.109 0.9318 1.383 99 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.89 0.863 0.878 0.706 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 53/80 12G 1.089 0.9139 1.363 103 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.911 0.863 0.88 0.708 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 54/80 12G 1.09 0.9126 1.366 100 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.897 0.863 0.88 0.711 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 55/80 12G 1.088 0.9111 1.37 78 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.91 0.858 0.881 0.714 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 56/80 12G 1.083 0.8995 1.367 68 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.909 0.863 0.88 0.714 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 57/80 12G 1.083 0.8901 1.36 76 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.908 0.864 0.883 0.71 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 58/80 12G 1.074 0.8867 1.356 61 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.911 0.864 0.881 0.713 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 59/80 12G 1.076 0.885 1.358 108 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.907 0.866 0.885 0.714 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 60/80 12G 1.087 0.9013 1.363 117 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.911 0.859 0.885 0.714 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 61/80 12G 1.066 0.8711 1.355 69 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.91 0.863 0.89 0.714 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 62/80 12G 1.069 0.8776 1.346 66 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.917 0.862 0.885 0.717 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 63/80 12G 1.053 0.8538 1.343 80 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.918 0.862 0.888 0.718 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 64/80 12G 1.053 0.8461 1.334 58 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.911 0.864 0.887 0.718 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 65/80 12G 1.045 0.8357 1.332 86 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.909 0.867 0.885 0.718 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 66/80 12G 1.046 0.8335 1.333 128 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.908 0.867 0.887 0.72 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 67/80 12G 1.025 0.818 1.319 53 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.916 0.865 0.887 0.717 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 68/80 12G 1.043 0.8391 1.333 97 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.907 0.864 0.885 0.719 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 69/80 12G 1.02 0.816 1.314 98 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.91 0.862 0.882 0.719 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 70/80 12G 1.026 0.8182 1.318 68 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.91 0.864 0.884 0.72 Closing dataloader mosaic albumentations: Blur(p=0.01, blur_limit=(3, 7)), MedianBlur(p=0.01, blur_limit=(3, 7)), ToGray(p=0.01, num_output_channels=3, method='weighted_average'), CLAHE(p=0.01, clip_limit=(1.0, 4.0), tile_grid_size=(8, 8)) /opt/conda/lib/python3.10/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock. self.pid = os.fork() Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 71/80 12G 0.9291 0.6012 1.243 45 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.906 0.865 0.885 0.719 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 72/80 11.9G 0.9103 0.5719 1.231 34 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.903 0.867 0.887 0.721 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 73/80 12G 0.9051 0.5584 1.22 37 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.913 0.865 0.885 0.721 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 74/80 11.9G 0.8919 0.5476 1.215 39 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.909 0.864 0.885 0.722 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 75/80 12G 0.8817 0.5305 1.207 36 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.908 0.862 0.885 0.724 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 76/80 12G 0.8784 0.5278 1.199 33 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.91 0.862 0.885 0.725 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 77/80 12G 0.8746 0.5236 1.199 49 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.911 0.861 0.886 0.725 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 78/80 12G 0.8674 0.5103 1.195 43 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.913 0.864 0.885 0.724 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 79/80 11.9G 0.8655 0.5087 1.193 39 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.908 0.864 0.886 0.725 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 80/80 12G 0.8464 0.5037 1.179 49 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.91 0.862 0.887 0.725 80 epochs completed in 6.254 hours. Optimizer stripped from runs/detect/train/weights/last.pt, 18.9MB Optimizer stripped from runs/detect/train/weights/best.pt, 18.9MB Validating runs/detect/train/weights/best.pt... Ultralytics 8.3.78 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB) YOLOv12s summary (fused): 159 layers, 9,231,654 parameters, 0 gradients, 21.2 GFLOPs Class Images Instances Box(P R mAP50 m all 1062 2081 0.894 0.869 0.879 0.718 Fire 859 1320 0.913 0.939 0.931 0.766 Smoke 559 761 0.875 0.8 0.828 0.669 Speed: 0.2ms preprocess, 15.5ms inference, 0.0ms loss, 0.8ms postprocess per image Results saved to runs/detect/train π‘ Learn more at https://docs.ultralytics.com/modes/train
InΒ [4]:
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
img = mpimg.imread('/kaggle/working/runs/detect/train/results.png')
imgplot = plt.imshow(img)
plt.show()
InΒ [5]:
img = mpimg.imread('/kaggle/working/runs/detect/train/confusion_matrix_normalized.png')
imgplot = plt.imshow(img)
plt.show()
InΒ [6]:
img = mpimg.imread('/kaggle/working/runs/detect/train/P_curve.png')
imgplot = plt.imshow(img)
plt.show()
InΒ [7]:
img = mpimg.imread('/kaggle/working/runs/detect/train/val_batch0_pred.jpg')
imgplot = plt.imshow(img)
plt.show()
InΒ [Β ]:
!yolo task=detect mode=predict model="/kaggle/working/runs/detect/train/weights/best.pt" device=0 source="/kaggle/working/smokeandfire-2/test/images" save=True
InΒ [9]:
from glob import glob
from PIL import Image
import numpy as np
img_path = glob('/kaggle/working/runs/detect/predict/*.jpg')
img_path = np.random.choice(img_path, 10)
for image_path in img_path:
plt.imshow(Image.open(image_path))
plt.axis("off")
plt.show()
print("\n")
InΒ [11]:
!yolo predict model="/kaggle/working/runs/detect/train/weights/best.pt", source="/kaggle/input/fireimage/fire.109.jpg"
WARNING β οΈ argument 'model=/kaggle/working/runs/detect/train/weights/best.pt,' does not require trailing comma ',', updating to 'model=/kaggle/working/runs/detect/train/weights/best.pt'.
Ultralytics 8.3.78 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB)
YOLOv12s summary (fused): 159 layers, 9,231,654 parameters, 0 gradients, 21.2 GFLOPs
image 1/1 /kaggle/input/fireimage/fire.109.jpg: 384x640 6 Fires, 1 Smoke, 65.4ms
Speed: 6.0ms preprocess, 65.4ms inference, 160.2ms postprocess per image at shape (1, 3, 384, 640)
Results saved to runs/detect/predict2
π‘ Learn more at https://docs.ultralytics.com/modes/predict
InΒ [12]:
# Path to the output image
output_image_path = "/kaggle/working/runs/detect/predict2/fire.109.jpg" # Adjust if the file name is different
output_image = cv2.imread(output_image_path)
output_image = cv2.cvtColor(output_image, cv2.COLOR_BGR2RGB) # Convert BGR to RGB for correct display
# Display the output image
plt.figure(figsize=(8, 8))
plt.imshow(output_image)
plt.title('Detected Image - fire.109.jpg')
plt.axis('off')
plt.show()
InΒ [15]:
!yolo predict model="/kaggle/working/runs/detect/train/weights/best.pt", source="/kaggle/input/fireeu/images.jpg"
WARNING β οΈ argument 'model=/kaggle/working/runs/detect/train/weights/best.pt,' does not require trailing comma ',', updating to 'model=/kaggle/working/runs/detect/train/weights/best.pt'.
Ultralytics 8.3.78 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB)
YOLOv12s summary (fused): 159 layers, 9,231,654 parameters, 0 gradients, 21.2 GFLOPs
image 1/1 /kaggle/input/fireeu/images.jpg: 640x640 1 Fire, 2 Smokes, 14.9ms
Speed: 7.2ms preprocess, 14.9ms inference, 165.9ms postprocess per image at shape (1, 3, 640, 640)
Results saved to runs/detect/predict4
π‘ Learn more at https://docs.ultralytics.com/modes/predict
InΒ [16]:
import cv2
import matplotlib.pyplot as plt
# Path to the output image
output_image_path = "/kaggle/working/runs/detect/predict4/images.jpg" # Adjust if the file name is different
# Load the output image
output_image = cv2.imread(output_image_path)
output_image = cv2.cvtColor(output_image, cv2.COLOR_BGR2RGB) # Convert BGR to RGB for correct display
# Display the output image
plt.figure(figsize=(8, 8))
plt.imshow(output_image)
plt.title('Detected Image - fire.109.jpg')
plt.axis('off') # Hide axes
plt.show()
yolo v8 small with 80 epochΒΆ
InΒ [17]:
!yolo train model=yolov8s.pt data='/kaggle/working/smokeandfire-2/data.yaml' epochs=80 imgsz=640 batch=32 augment=True lr0=0.005 lrf=0.01 weight_decay=0.0005 optimizer=SGD mosaic=True hsv_h=0.015 hsv_s=0.7 hsv_v=0.4 degrees=15 translate=0.1 scale=0.5 shear=0.2 flipud=0.5 fliplr=0.5 mixup=0.2
Downloading https://github.com/ultralytics/assets/releases/download/v8.3.0/yolov8s.pt to 'yolov8s.pt'... 100%|ββββββββββββββββββββββββββββββββββββββ| 21.5M/21.5M [00:00<00:00, 42.8MB/s] Ultralytics 8.3.78 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB) engine/trainer: task=detect, mode=train, model=yolov8s.pt, data=/kaggle/working/smokeandfire-2/data.yaml, epochs=80, time=None, patience=100, batch=32, imgsz=640, save=True, save_period=-1, cache=False, device=None, workers=8, project=None, name=train2, exist_ok=False, pretrained=True, optimizer=SGD, verbose=True, seed=0, deterministic=True, single_cls=False, rect=False, cos_lr=False, close_mosaic=10, resume=False, amp=True, fraction=1.0, profile=False, freeze=None, multi_scale=False, overlap_mask=True, mask_ratio=4, dropout=0.0, val=True, split=val, save_json=False, save_hybrid=False, conf=None, iou=0.7, max_det=300, half=False, dnn=False, plots=True, source=None, vid_stride=1, stream_buffer=False, visualize=False, augment=True, agnostic_nms=False, classes=None, retina_masks=False, embed=None, show=False, save_frames=False, save_txt=False, save_conf=False, save_crop=False, show_labels=True, show_conf=True, show_boxes=True, line_width=None, format=torchscript, keras=False, optimize=False, int8=False, dynamic=False, simplify=True, opset=None, workspace=None, nms=False, lr0=0.005, lrf=0.01, momentum=0.937, weight_decay=0.0005, warmup_epochs=3.0, warmup_momentum=0.8, warmup_bias_lr=0.1, box=7.5, cls=0.5, dfl=1.5, pose=12.0, kobj=1.0, nbs=64, hsv_h=0.015, hsv_s=0.7, hsv_v=0.4, degrees=15, translate=0.1, scale=0.5, shear=0.2, perspective=0.0, flipud=0.5, fliplr=0.5, bgr=0.0, mosaic=True, mixup=0.2, copy_paste=0.0, copy_paste_mode=flip, auto_augment=randaugment, erasing=0.4, crop_fraction=1.0, cfg=None, tracker=botsort.yaml, save_dir=runs/detect/train2 Overriding model.yaml nc=80 with nc=2 from n params module arguments 0 -1 1 928 ultralytics.nn.modules.conv.Conv [3, 32, 3, 2] 1 -1 1 18560 ultralytics.nn.modules.conv.Conv [32, 64, 3, 2] 2 -1 1 29056 ultralytics.nn.modules.block.C2f [64, 64, 1, True] 3 -1 1 73984 ultralytics.nn.modules.conv.Conv [64, 128, 3, 2] 4 -1 2 197632 ultralytics.nn.modules.block.C2f [128, 128, 2, True] 5 -1 1 295424 ultralytics.nn.modules.conv.Conv [128, 256, 3, 2] 6 -1 2 788480 ultralytics.nn.modules.block.C2f [256, 256, 2, True] 7 -1 1 1180672 ultralytics.nn.modules.conv.Conv [256, 512, 3, 2] 8 -1 1 1838080 ultralytics.nn.modules.block.C2f [512, 512, 1, True] 9 -1 1 656896 ultralytics.nn.modules.block.SPPF [512, 512, 5] 10 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 11 [-1, 6] 1 0 ultralytics.nn.modules.conv.Concat [1] 12 -1 1 591360 ultralytics.nn.modules.block.C2f [768, 256, 1] 13 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 14 [-1, 4] 1 0 ultralytics.nn.modules.conv.Concat [1] 15 -1 1 148224 ultralytics.nn.modules.block.C2f [384, 128, 1] 16 -1 1 147712 ultralytics.nn.modules.conv.Conv [128, 128, 3, 2] 17 [-1, 12] 1 0 ultralytics.nn.modules.conv.Concat [1] 18 -1 1 493056 ultralytics.nn.modules.block.C2f [384, 256, 1] 19 -1 1 590336 ultralytics.nn.modules.conv.Conv [256, 256, 3, 2] 20 [-1, 9] 1 0 ultralytics.nn.modules.conv.Concat [1] 21 -1 1 1969152 ultralytics.nn.modules.block.C2f [768, 512, 1] 22 [15, 18, 21] 1 2116822 ultralytics.nn.modules.head.Detect [2, [128, 256, 512]] Model summary: 129 layers, 11,136,374 parameters, 11,136,358 gradients, 28.6 GFLOPs Transferred 349/355 items from pretrained weights TensorBoard: Start with 'tensorboard --logdir runs/detect/train2', view at http://localhost:6006/ Freezing layer 'model.22.dfl.conv.weight' AMP: running Automatic Mixed Precision (AMP) checks... AMP: checks passed β train: Scanning /kaggle/working/smokeandfire-2/train/labels.cache... 8243 images WARNING β οΈ Box and segment counts should be equal, but got len(segments) = 12, len(boxes) = 16365. To resolve this only boxes will be used and all segments will be removed. To avoid this please supply either a detect or segment dataset, not a detect-segment mixed dataset. /opt/conda/lib/python3.10/site-packages/albumentations/__init__.py:24: UserWarning: A new version of Albumentations is available: 2.0.4 (you have 1.4.21). Upgrade using: pip install -U albumentations. To disable automatic update checks, set the environment variable NO_ALBUMENTATIONS_UPDATE to 1. check_for_updates() albumentations: Blur(p=0.01, blur_limit=(3, 7)), MedianBlur(p=0.01, blur_limit=(3, 7)), ToGray(p=0.01, num_output_channels=3, method='weighted_average'), CLAHE(p=0.01, clip_limit=(1.0, 4.0), tile_grid_size=(8, 8)) /opt/conda/lib/python3.10/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock. self.pid = os.fork() val: Scanning /kaggle/working/smokeandfire-2/valid/labels.cache... 1062 images, /opt/conda/lib/python3.10/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock. self.pid = os.fork() Plotting labels to runs/detect/train2/labels.jpg... optimizer: SGD(lr=0.005, momentum=0.937) with parameter groups 57 weight(decay=0.0), 64 weight(decay=0.0005), 63 bias(decay=0.0) TensorBoard: model graph visualization added β Image sizes 640 train, 640 val Using 4 dataloader workers Logging results to runs/detect/train2 Starting training for 80 epochs... Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 1/80 7.49G 1.777 2.502 1.856 96 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.68 0.54 0.589 0.424 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 2/80 7.68G 1.449 1.589 1.595 76 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.817 0.742 0.776 0.552 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 3/80 7.72G 1.443 1.496 1.573 53 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.794 0.726 0.759 0.537 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 4/80 7.72G 1.466 1.524 1.597 90 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.82 0.734 0.767 0.516 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 5/80 7.72G 1.449 1.478 1.577 95 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.824 0.751 0.792 0.576 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 6/80 7.79G 1.418 1.43 1.568 62 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.856 0.783 0.809 0.57 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 7/80 7.8G 1.397 1.393 1.543 72 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.862 0.783 0.813 0.573 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 8/80 7.71G 1.381 1.369 1.543 82 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.844 0.808 0.824 0.598 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 9/80 7.71G 1.361 1.349 1.527 54 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.857 0.779 0.811 0.597 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 10/80 7.71G 1.344 1.316 1.508 91 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.868 0.801 0.826 0.595 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 11/80 7.71G 1.327 1.293 1.498 88 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.869 0.806 0.829 0.608 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 12/80 7.73G 1.323 1.276 1.496 112 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.867 0.805 0.83 0.619 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 13/80 7.73G 1.314 1.262 1.485 76 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.857 0.795 0.791 0.585 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 14/80 7.71G 1.301 1.243 1.48 80 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.878 0.817 0.845 0.628 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 15/80 7.79G 1.297 1.242 1.476 114 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.877 0.814 0.84 0.63 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 16/80 7.79G 1.279 1.228 1.465 91 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.876 0.822 0.843 0.626 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 17/80 7.79G 1.269 1.212 1.465 65 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.876 0.835 0.852 0.638 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 18/80 7.79G 1.271 1.204 1.467 137 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.882 0.821 0.84 0.637 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 19/80 7.79G 1.267 1.19 1.46 92 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.864 0.83 0.842 0.629 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 20/80 7.71G 1.262 1.182 1.453 92 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.87 0.829 0.844 0.639 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 21/80 7.72G 1.241 1.152 1.433 75 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.886 0.847 0.849 0.644 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 22/80 7.79G 1.252 1.166 1.444 87 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.871 0.835 0.849 0.646 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 23/80 7.79G 1.231 1.145 1.435 117 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.892 0.843 0.863 0.658 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 24/80 7.79G 1.232 1.143 1.43 129 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.881 0.847 0.856 0.658 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 25/80 7.73G 1.224 1.12 1.416 80 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.836 0.857 0.66 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 26/80 7.73G 1.226 1.128 1.419 78 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.89 0.843 0.859 0.658 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 27/80 7.72G 1.206 1.103 1.41 75 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.882 0.847 0.861 0.663 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 28/80 7.79G 1.217 1.106 1.416 71 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.897 0.847 0.867 0.668 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 29/80 7.72G 1.206 1.099 1.407 71 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.884 0.849 0.861 0.667 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 30/80 7.79G 1.2 1.086 1.401 70 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.896 0.855 0.861 0.667 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 31/80 7.72G 1.186 1.069 1.388 68 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.897 0.838 0.856 0.672 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 32/80 7.8G 1.189 1.083 1.396 75 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.9 0.852 0.857 0.665 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 33/80 7.81G 1.186 1.069 1.393 97 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.901 0.845 0.86 0.67 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 34/80 7.73G 1.188 1.055 1.389 72 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.902 0.849 0.858 0.674 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 35/80 7.79G 1.173 1.063 1.389 72 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.894 0.845 0.864 0.687 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 36/80 7.8G 1.152 1.028 1.366 100 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.888 0.851 0.858 0.673 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 37/80 7.8G 1.163 1.023 1.374 81 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.893 0.851 0.861 0.679 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 38/80 7.72G 1.17 1.042 1.383 74 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.896 0.844 0.862 0.677 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 39/80 7.73G 1.153 1.022 1.368 98 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.891 0.847 0.861 0.674 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 40/80 7.79G 1.144 1.009 1.361 77 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.899 0.841 0.858 0.679 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 41/80 7.72G 1.153 1.011 1.367 78 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.899 0.85 0.859 0.69 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 42/80 7.72G 1.141 0.9984 1.358 111 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.904 0.852 0.861 0.687 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 43/80 7.72G 1.148 1.001 1.352 111 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.891 0.857 0.857 0.685 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 44/80 7.8G 1.131 0.9783 1.351 67 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.898 0.854 0.868 0.691 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 45/80 7.72G 1.134 0.9825 1.356 79 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.904 0.854 0.868 0.692 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 46/80 7.72G 1.124 0.9722 1.342 94 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.896 0.855 0.871 0.697 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 47/80 7.72G 1.112 0.9707 1.334 77 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.899 0.864 0.872 0.698 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 48/80 7.72G 1.127 0.9679 1.342 57 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.9 0.86 0.866 0.696 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 49/80 7.72G 1.11 0.9525 1.333 96 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.898 0.854 0.869 0.698 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 50/80 7.72G 1.109 0.9496 1.332 91 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.897 0.86 0.874 0.702 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 51/80 7.72G 1.106 0.9419 1.333 76 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.901 0.863 0.873 0.701 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 52/80 7.79G 1.109 0.9338 1.332 99 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.905 0.863 0.87 0.7 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 53/80 7.8G 1.084 0.9181 1.313 103 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.907 0.856 0.87 0.696 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 54/80 7.71G 1.087 0.9224 1.312 100 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.894 0.866 0.871 0.696 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 55/80 7.8G 1.09 0.9145 1.313 78 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.904 0.863 0.875 0.705 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 56/80 7.79G 1.078 0.9079 1.31 68 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.901 0.868 0.871 0.704 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 57/80 7.8G 1.078 0.905 1.308 76 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.903 0.863 0.877 0.707 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 58/80 7.8G 1.073 0.8991 1.307 61 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.898 0.867 0.871 0.703 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 59/80 7.72G 1.077 0.8967 1.31 108 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.905 0.865 0.873 0.703 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 60/80 7.72G 1.087 0.9122 1.313 117 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.914 0.865 0.875 0.703 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 61/80 7.8G 1.069 0.8888 1.3 69 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.912 0.864 0.877 0.708 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 62/80 7.71G 1.069 0.8887 1.293 66 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.903 0.869 0.875 0.708 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 63/80 7.72G 1.053 0.8714 1.29 80 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.907 0.866 0.873 0.706 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 64/80 7.71G 1.051 0.8612 1.282 58 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.909 0.868 0.875 0.71 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 65/80 7.72G 1.046 0.8514 1.275 86 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.91 0.867 0.875 0.712 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 66/80 7.8G 1.043 0.8504 1.282 128 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.907 0.872 0.877 0.71 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 67/80 7.8G 1.022 0.8309 1.267 53 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.908 0.865 0.875 0.709 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 68/80 7.79G 1.043 0.8582 1.286 97 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.905 0.862 0.873 0.707 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 69/80 7.72G 1.028 0.8307 1.271 98 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.908 0.865 0.871 0.708 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 70/80 7.8G 1.027 0.8349 1.271 68 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.906 0.864 0.872 0.706 Closing dataloader mosaic albumentations: Blur(p=0.01, blur_limit=(3, 7)), MedianBlur(p=0.01, blur_limit=(3, 7)), ToGray(p=0.01, num_output_channels=3, method='weighted_average'), CLAHE(p=0.01, clip_limit=(1.0, 4.0), tile_grid_size=(8, 8)) /opt/conda/lib/python3.10/multiprocessing/popen_fork.py:66: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock. self.pid = os.fork() Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 71/80 7.72G 0.9332 0.6394 1.202 45 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.911 0.855 0.87 0.705 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 72/80 7.8G 0.9181 0.6044 1.191 34 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.909 0.858 0.874 0.708 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 73/80 7.72G 0.9109 0.59 1.184 37 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.903 0.863 0.874 0.708 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 74/80 7.72G 0.8956 0.5777 1.175 39 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.909 0.857 0.874 0.711 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 75/80 7.72G 0.8891 0.5623 1.173 36 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.906 0.861 0.875 0.713 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 76/80 7.72G 0.8886 0.5596 1.168 33 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.9 0.864 0.876 0.714 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 77/80 7.72G 0.8827 0.5559 1.16 49 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.903 0.86 0.874 0.714 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 78/80 7.8G 0.8754 0.5481 1.159 43 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.902 0.862 0.874 0.715 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 79/80 7.71G 0.8745 0.5433 1.157 39 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.902 0.863 0.874 0.714 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 80/80 7.72G 0.8573 0.5318 1.147 49 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.903 0.863 0.875 0.716 80 epochs completed in 3.188 hours. Optimizer stripped from runs/detect/train2/weights/last.pt, 22.5MB Optimizer stripped from runs/detect/train2/weights/best.pt, 22.5MB Validating runs/detect/train2/weights/best.pt... Ultralytics 8.3.78 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB) Model summary (fused): 72 layers, 11,126,358 parameters, 0 gradients, 28.4 GFLOPs Class Images Instances Box(P R mAP50 m all 1062 2081 0.895 0.866 0.878 0.712 Fire 859 1320 0.912 0.94 0.938 0.764 Smoke 559 761 0.879 0.791 0.819 0.661 Speed: 0.1ms preprocess, 10.0ms inference, 0.0ms loss, 0.8ms postprocess per image Results saved to runs/detect/train2 π‘ Learn more at https://docs.ultralytics.com/modes/train
test our final model yolo v8ΒΆ
InΒ [9]:
!yolo val model=/kaggle/input/best/pytorch/default/1/bestt.pt data=/kaggle/working/smokeandfire-2/data.yaml split=train
Ultralytics 8.3.91 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB) Model summary (fused): 72 layers, 11,126,358 parameters, 0 gradients, 28.4 GFLOPs val: Scanning /kaggle/working/smokeandfire-2/train/labels... 8243 images, 11 bac val: New cache created: /kaggle/working/smokeandfire-2/train/labels.cache WARNING β οΈ Box and segment counts should be equal, but got len(segments) = 12, len(boxes) = 16365. To resolve this only boxes will be used and all segments will be removed. To avoid this please supply either a detect or segment dataset, not a detect-segment mixed dataset. Class Images Instances Box(P R mAP50 m all 8243 16365 0.925 0.876 0.935 0.709 Fire 6693 11239 0.928 0.898 0.944 0.709 Smoke 3785 5126 0.921 0.854 0.927 0.708 Speed: 0.2ms preprocess, 4.6ms inference, 0.0ms loss, 0.8ms postprocess per image Results saved to runs/detect/val3 π‘ Learn more at https://docs.ultralytics.com/modes/val
InΒ [10]:
!yolo val model=/kaggle/input/best/pytorch/default/1/bestt.pt data=/kaggle/working/smokeandfire-2/data.yaml split=test
Ultralytics 8.3.91 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB) Model summary (fused): 72 layers, 11,126,358 parameters, 0 gradients, 28.4 GFLOPs val: Scanning /kaggle/working/smokeandfire-2/test/labels... 543 images, 1 backgr val: New cache created: /kaggle/working/smokeandfire-2/test/labels.cache Class Images Instances Box(P R mAP50 m all 543 1065 0.943 0.881 0.902 0.788 Fire 528 748 0.965 0.951 0.963 0.819 Smoke 190 317 0.921 0.811 0.841 0.757 Speed: 0.5ms preprocess, 4.4ms inference, 0.0ms loss, 1.0ms postprocess per image Results saved to runs/detect/val4 π‘ Learn more at https://docs.ultralytics.com/modes/val
InΒ [Β ]:
#pip install ultralytics==8.3.48
InΒ [Β ]:
#pip install torch==2.4.0 torchvision torchaudio
try to combine different sourcesΒΆ
InΒ [Β ]:
# merge
InΒ [Β ]:
!pip install roboflow
from roboflow import Roboflow
rf = Roboflow(api_key="GZBlBPlRdRvB6uR7BqaS")
project = rf.workspace("fire-and-smoke-detection").project("fds-p1vmt")
version = project.version(1)
dataset = version.download("yolov8")
InΒ [Β ]:
!pip install roboflow #Ψ―Ω
from roboflow import Roboflow
rf = Roboflow(api_key="GZBlBPlRdRvB6uR7BqaS")
project = rf.workspace("comsats-university-islamabad-wah-campus").project("smoke-detection-eopca")
version = project.version(1)
dataset = version.download("yolov8")
InΒ [Β ]:
!pip install roboflow
from roboflow import Roboflow
rf = Roboflow(api_key="GZBlBPlRdRvB6uR7BqaS")
project = rf.workspace("smoke-detection").project("smoke100-uwe4t")
version = project.version(5)
dataset = version.download("yolov8")
InΒ [6]:
import os
import shutil
import yaml
from glob import glob
import random
dataset1 = "/kaggle/working/FDS-1"
dataset2 = "/kaggle/working/smokeandfire-2"
dataset3 = "/kaggle/working/Smoke-Detection-1"
dataset4 = "/kaggle/working/smoke-upm-1"
dataset5 = "/kaggle/working/Smoke100-5"
output_dir = "/kaggle/working/combined_dataset_final6"
for split in ['train', 'valid', 'test']:
for sub in ['images', 'labels']:
os.makedirs(os.path.join(output_dir, split, sub), exist_ok=True)
def load_yaml(path):
with open(path, 'r') as f:
return yaml.safe_load(f)
data1 = load_yaml(os.path.join(dataset1, 'data.yaml'))
data2 = load_yaml(os.path.join(dataset2, 'data.yaml'))
data3 = load_yaml(os.path.join(dataset3, 'data.yaml'))
data1['names'] = [name.lower() for name in data1['names']]
data2['names'] = [name.lower() for name in data2['names']]
data3['names'] = [name.lower() for name in data3['names']]
combined_classes = sorted(set(data1['names'] + data2['names'] + data3['names'] + ['smoke'])) # β¨ Add smoke manually
name_to_index = {name: i for i, name in enumerate(combined_classes)}
# Helper to copy and update label
def copy_dataset(src_root, split, src_data):
img_dir = os.path.join(src_root, split, 'images')
lbl_dir = os.path.join(src_root, split, 'labels')
for img_path in glob(f"{img_dir}/*.jpg") + glob(f"{img_dir}/*.png"):
base = os.path.splitext(os.path.basename(img_path))[0]
label_path = os.path.join(lbl_dir, f"{base}.txt")
# Copy image
shutil.copy(img_path, os.path.join(output_dir, split, 'images'))
if os.path.exists(label_path):
with open(label_path, 'r') as f:
lines = f.readlines()
new_lines = []
for line in lines:
parts = line.strip().split()
old_idx = int(parts[0])
cls_name = src_data['names'][old_idx]
new_idx = name_to_index[cls_name]
new_line = " ".join([str(new_idx)] + parts[1:])
new_lines.append(new_line)
with open(os.path.join(output_dir, split, 'labels', f"{base}.txt"), 'w') as f:
f.write("\n".join(new_lines))
# Helper to copy datasets
def copy_smoke_dataset(src_root, split):
img_dir = os.path.join(src_root, split, 'images')
lbl_dir = os.path.join(src_root, split, 'labels')
for img_path in glob(f"{img_dir}/*.jpg") + glob(f"{img_dir}/*.png"):
base = os.path.splitext(os.path.basename(img_path))[0]
label_path = os.path.join(lbl_dir, f"{base}.txt")
# Copy image
shutil.copy(img_path, os.path.join(output_dir, split, 'images'))
if os.path.exists(label_path):
with open(label_path, 'r') as f:
lines = f.readlines()
smoke_idx = name_to_index['smoke']
new_lines = []
for line in lines:
parts = line.strip().split()
new_line = " ".join([str(smoke_idx)] + parts[1:])
new_lines.append(new_line)
with open(os.path.join(output_dir, split, 'labels', f"{base}.txt"), 'w') as f:
f.write("\n".join(new_lines))
def copy_limited_smoke_dataset(src_root, max_images=4200):
img_dir = os.path.join(src_root, 'train', 'images')
lbl_dir = os.path.join(src_root, 'train', 'labels')
images = glob(f"{img_dir}/*.jpg") + glob(f"{img_dir}/*.png")
images = random.sample(images, min(max_images, len(images)))
for img_path in images:
base = os.path.splitext(os.path.basename(img_path))[0]
label_path = os.path.join(lbl_dir, f"{base}.txt")
# Copy image
shutil.copy(img_path, os.path.join(output_dir, 'train', 'images'))
if os.path.exists(label_path):
with open(label_path, 'r') as f:
lines = f.readlines()
smoke_idx = name_to_index['smoke']
new_lines = []
for line in lines:
parts = line.strip().split()
new_line = " ".join([str(smoke_idx)] + parts[1:])
new_lines.append(new_line)
with open(os.path.join(output_dir, 'train', 'labels', f"{base}.txt"), 'w') as f:
f.write("\n".join(new_lines))
for split in ['train', 'valid', 'test']:
copy_dataset(dataset1, split, data1)
copy_dataset(dataset2, split, data2)
copy_dataset(dataset3, split, data3)
copy_smoke_dataset(dataset4, split)
copy_limited_smoke_dataset(dataset5, max_images=4200)
new_yaml = {
'train': os.path.join(output_dir, 'train', 'images'),
'val': os.path.join(output_dir, 'valid', 'images'),
'test': os.path.join(output_dir, 'test', 'images'),
'nc': len(combined_classes),
'names': combined_classes
}
with open(os.path.join(output_dir, 'data.yaml'), 'w') as f:
yaml.dump(new_yaml, f)
print(" merged")
merged
InΒ [7]:
from collections import defaultdict
splits = ['train', 'valid', 'test']
labels_root = "/kaggle/working/combined_dataset_final6"
class_names = combined_classes
for split in splits:
class_counts = defaultdict(int)
label_files = glob(os.path.join(labels_root, split, 'labels', '*.txt'))
for label_file in label_files:
with open(label_file, 'r') as f:
lines = f.readlines()
for line in lines:
cls_id = int(line.strip().split()[0])
class_counts[cls_id] += 1
print(f"\n Split: {split.upper()}")
total_labels = 0
for cls_id, count in class_counts.items():
print(f" - {class_names[cls_id]}: {count} images")
total_labels += count
print(f" Total labels in {split}: {total_labels}")
Split: TRAIN - smoke: 15483 images - fire: 16051 images Total labels in train: 31534 Split: VALID - fire: 2708 images - smoke: 2566 images Total labels in valid: 5274 Split: TEST - fire: 1341 images - smoke: 1104 images Total labels in test: 2445
InΒ [8]:
!yolo train model=yolov8s.pt data='/kaggle/working/combined_dataset_final6/data.yaml' epochs=80 imgsz=640 batch=32 augment=True lr0=0.005 lrf=0.01 weight_decay=0.0005 optimizer=SGD mosaic=True hsv_h=0.015 hsv_s=0.7 hsv_v=0.4 degrees=15 translate=0.1 scale=0.5 shear=0.2 flipud=0.5 fliplr=0.5 mixup=0.2
Downloading https://github.com/ultralytics/assets/releases/download/v8.3.0/yolov8s.pt to 'yolov8s.pt'... 100%|βββββββββββββββββββββββββββββββββββββββ| 21.5M/21.5M [00:00<00:00, 331MB/s] Ultralytics 8.3.126 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB) engine/trainer: agnostic_nms=False, amp=True, augment=True, auto_augment=randaugment, batch=32, bgr=0.0, box=7.5, cache=False, cfg=None, classes=None, close_mosaic=10, cls=0.5, conf=None, copy_paste=0.0, copy_paste_mode=flip, cos_lr=False, cutmix=0.0, data=/kaggle/working/combined_dataset_final6/data.yaml, degrees=15, deterministic=True, device=cuda:0, dfl=1.5, dnn=False, dropout=0.0, dynamic=False, embed=None, epochs=80, erasing=0.4, exist_ok=False, fliplr=0.5, flipud=0.5, format=torchscript, fraction=1.0, freeze=None, half=False, hsv_h=0.015, hsv_s=0.7, hsv_v=0.4, imgsz=640, int8=False, iou=0.7, keras=False, kobj=1.0, line_width=None, lr0=0.005, lrf=0.01, mask_ratio=4, max_det=300, mixup=0.2, mode=train, model=yolov8s.pt, momentum=0.937, mosaic=True, multi_scale=False, name=train, nbs=64, nms=False, opset=None, optimize=False, optimizer=SGD, overlap_mask=True, patience=100, perspective=0.0, plots=True, pose=12.0, pretrained=True, profile=False, project=None, rect=False, resume=False, retina_masks=False, save=True, save_conf=False, save_crop=False, save_dir=runs/detect/train, save_frames=False, save_json=False, save_period=-1, save_txt=False, scale=0.5, seed=0, shear=0.2, show=False, show_boxes=True, show_conf=True, show_labels=True, simplify=True, single_cls=False, source=None, split=val, stream_buffer=False, task=detect, time=None, tracker=botsort.yaml, translate=0.1, val=True, verbose=True, vid_stride=1, visualize=False, warmup_bias_lr=0.1, warmup_epochs=3.0, warmup_momentum=0.8, weight_decay=0.0005, workers=8, workspace=None Downloading https://ultralytics.com/assets/Arial.ttf to '/root/.config/Ultralytics/Arial.ttf'... 100%|ββββββββββββββββββββββββββββββββββββββββ| 755k/755k [00:00<00:00, 42.7MB/s] Overriding model.yaml nc=80 with nc=2 from n params module arguments 0 -1 1 928 ultralytics.nn.modules.conv.Conv [3, 32, 3, 2] 1 -1 1 18560 ultralytics.nn.modules.conv.Conv [32, 64, 3, 2] 2 -1 1 29056 ultralytics.nn.modules.block.C2f [64, 64, 1, True] 3 -1 1 73984 ultralytics.nn.modules.conv.Conv [64, 128, 3, 2] 4 -1 2 197632 ultralytics.nn.modules.block.C2f [128, 128, 2, True] 5 -1 1 295424 ultralytics.nn.modules.conv.Conv [128, 256, 3, 2] 6 -1 2 788480 ultralytics.nn.modules.block.C2f [256, 256, 2, True] 7 -1 1 1180672 ultralytics.nn.modules.conv.Conv [256, 512, 3, 2] 8 -1 1 1838080 ultralytics.nn.modules.block.C2f [512, 512, 1, True] 9 -1 1 656896 ultralytics.nn.modules.block.SPPF [512, 512, 5] 10 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 11 [-1, 6] 1 0 ultralytics.nn.modules.conv.Concat [1] 12 -1 1 591360 ultralytics.nn.modules.block.C2f [768, 256, 1] 13 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 14 [-1, 4] 1 0 ultralytics.nn.modules.conv.Concat [1] 15 -1 1 148224 ultralytics.nn.modules.block.C2f [384, 128, 1] 16 -1 1 147712 ultralytics.nn.modules.conv.Conv [128, 128, 3, 2] 17 [-1, 12] 1 0 ultralytics.nn.modules.conv.Concat [1] 18 -1 1 493056 ultralytics.nn.modules.block.C2f [384, 256, 1] 19 -1 1 590336 ultralytics.nn.modules.conv.Conv [256, 256, 3, 2] 20 [-1, 9] 1 0 ultralytics.nn.modules.conv.Concat [1] 21 -1 1 1969152 ultralytics.nn.modules.block.C2f [768, 512, 1] 22 [15, 18, 21] 1 2116822 ultralytics.nn.modules.head.Detect [2, [128, 256, 512]] Model summary: 129 layers, 11,136,374 parameters, 11,136,358 gradients, 28.6 GFLOPs Transferred 349/355 items from pretrained weights Freezing layer 'model.22.dfl.conv.weight' AMP: running Automatic Mixed Precision (AMP) checks... Downloading https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11n.pt to 'yolo11n.pt'... 100%|βββββββββββββββββββββββββββββββββββββββ| 5.35M/5.35M [00:00<00:00, 198MB/s] AMP: checks passed β train: Fast image access β (ping: 0.0Β±0.0 ms, read: 780.5Β±596.0 MB/s, size: 23.4 KB) train: Scanning /kaggle/working/combined_dataset_final6/train/labels... 21725 im train: New cache created: /kaggle/working/combined_dataset_final6/train/labels.cache WARNING β οΈ Box and segment counts should be equal, but got len(segments) = 22, len(boxes) = 31534. To resolve this only boxes will be used and all segments will be removed. To avoid this please supply either a detect or segment dataset, not a detect-segment mixed dataset. albumentations: Blur(p=0.01, blur_limit=(3, 7)), MedianBlur(p=0.01, blur_limit=(3, 7)), ToGray(p=0.01, num_output_channels=3, method='weighted_average'), CLAHE(p=0.01, clip_limit=(1.0, 4.0), tile_grid_size=(8, 8)) val: Fast image access β (ping: 0.0Β±0.0 ms, read: 631.9Β±180.0 MB/s, size: 35.9 KB) val: Scanning /kaggle/working/combined_dataset_final6/valid/labels... 3718 image val: New cache created: /kaggle/working/combined_dataset_final6/valid/labels.cache Plotting labels to runs/detect/train/labels.jpg... optimizer: SGD(lr=0.005, momentum=0.937) with parameter groups 57 weight(decay=0.0), 64 weight(decay=0.0005), 63 bias(decay=0.0) Image sizes 640 train, 640 val Using 4 dataloader workers Logging results to runs/detect/train Starting training for 80 epochs... Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 1/80 7.62G 1.789 2.306 1.899 123 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.689 0.575 0.605 0.327 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 2/80 7.64G 1.504 1.489 1.639 98 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.82 0.747 0.806 0.447 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 3/80 7.65G 1.489 1.452 1.629 115 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.734 0.672 0.712 0.38 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 4/80 7.65G 1.522 1.48 1.65 114 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.763 0.738 0.778 0.436 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 5/80 7.65G 1.481 1.406 1.615 104 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.802 0.778 0.81 0.463 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 6/80 7.65G 1.444 1.347 1.594 105 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.831 0.8 0.839 0.498 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 7/80 7.65G 1.407 1.289 1.567 92 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.846 0.815 0.843 0.518 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 8/80 7.65G 1.407 1.281 1.567 92 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.852 0.827 0.859 0.523 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 9/80 7.65G 1.386 1.243 1.546 106 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.854 0.839 0.868 0.528 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 10/80 7.65G 1.368 1.217 1.535 109 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.874 0.845 0.876 0.536 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 11/80 7.65G 1.355 1.197 1.524 87 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.883 0.848 0.882 0.545 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 12/80 7.65G 1.34 1.173 1.509 89 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.889 0.854 0.887 0.556 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 13/80 7.65G 1.335 1.161 1.508 115 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.882 0.859 0.886 0.57 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 14/80 7.65G 1.314 1.134 1.491 93 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.889 0.865 0.894 0.567 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 15/80 7.65G 1.309 1.129 1.486 117 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.893 0.867 0.895 0.575 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 16/80 7.65G 1.296 1.113 1.477 111 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.893 0.866 0.896 0.585 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 17/80 7.65G 1.296 1.1 1.474 93 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.894 0.867 0.896 0.579 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 18/80 7.65G 1.291 1.093 1.466 95 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.894 0.877 0.904 0.596 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 19/80 7.65G 1.281 1.077 1.461 94 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.903 0.881 0.903 0.593 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 20/80 7.65G 1.269 1.072 1.456 116 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.906 0.883 0.905 0.599 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 21/80 7.65G 1.263 1.06 1.452 109 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.904 0.886 0.91 0.606 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 22/80 7.65G 1.264 1.054 1.449 110 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.899 0.888 0.907 0.604 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 23/80 7.65G 1.26 1.049 1.442 110 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.912 0.883 0.909 0.609 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 24/80 7.65G 1.24 1.027 1.431 80 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.921 0.883 0.914 0.617 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 25/80 7.65G 1.239 1.02 1.431 85 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.916 0.888 0.913 0.616 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 26/80 7.65G 1.235 1.015 1.427 117 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.916 0.889 0.912 0.62 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 27/80 7.65G 1.234 1.013 1.428 89 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.919 0.888 0.917 0.624 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 28/80 7.65G 1.223 1.001 1.416 97 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.922 0.892 0.915 0.625 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 29/80 7.65G 1.221 0.9965 1.418 88 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.922 0.89 0.918 0.63 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 30/80 7.65G 1.215 0.9863 1.411 95 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.927 0.891 0.919 0.632 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 31/80 7.65G 1.208 0.9772 1.403 106 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.923 0.89 0.917 0.63 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 32/80 7.65G 1.209 0.972 1.408 80 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.917 0.892 0.916 0.634 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 33/80 7.65G 1.199 0.9618 1.398 97 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.92 0.891 0.917 0.636 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 34/80 7.65G 1.202 0.9712 1.402 99 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.922 0.894 0.917 0.639 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 35/80 7.65G 1.192 0.9521 1.398 119 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.921 0.896 0.919 0.641 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 36/80 7.65G 1.187 0.947 1.392 133 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.92 0.895 0.92 0.642 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 37/80 7.65G 1.186 0.9395 1.386 82 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.921 0.894 0.919 0.641 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 38/80 7.65G 1.177 0.9359 1.384 83 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.922 0.891 0.918 0.644 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 39/80 7.65G 1.175 0.9357 1.385 116 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.922 0.893 0.918 0.646 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 40/80 7.65G 1.173 0.9277 1.383 98 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.921 0.895 0.919 0.648 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 41/80 7.65G 1.169 0.919 1.376 98 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.925 0.891 0.919 0.648 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 42/80 7.65G 1.16 0.9141 1.369 89 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.928 0.894 0.92 0.649 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 43/80 7.65G 1.164 0.9189 1.375 92 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.923 0.898 0.921 0.651 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 44/80 7.65G 1.154 0.9057 1.362 91 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.926 0.897 0.922 0.651 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 45/80 7.65G 1.147 0.8937 1.357 106 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.925 0.9 0.922 0.653 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 46/80 7.65G 1.147 0.8967 1.357 115 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.924 0.9 0.922 0.654 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 47/80 7.65G 1.146 0.8927 1.356 117 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.925 0.899 0.922 0.654 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 48/80 7.65G 1.136 0.8756 1.352 111 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.926 0.9 0.923 0.654 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 49/80 7.65G 1.137 0.8812 1.35 81 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.926 0.9 0.923 0.656 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 50/80 7.65G 1.139 0.8849 1.354 90 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.927 0.898 0.923 0.656 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 51/80 7.65G 1.127 0.875 1.346 79 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.923 0.9 0.923 0.656 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 52/80 7.65G 1.123 0.8545 1.337 93 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.925 0.899 0.923 0.657 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 53/80 7.65G 1.118 0.8563 1.337 96 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.927 0.897 0.923 0.657 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 54/80 7.65G 1.112 0.8555 1.339 91 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.924 0.903 0.923 0.658 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 55/80 7.65G 1.12 0.8603 1.341 91 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.927 0.9 0.923 0.658 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 56/80 7.65G 1.106 0.8399 1.334 107 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.927 0.9 0.923 0.658 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 57/80 7.65G 1.113 0.8484 1.336 92 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.926 0.902 0.923 0.659 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 58/80 7.65G 1.098 0.8355 1.323 121 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.925 0.903 0.924 0.659 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 59/80 7.65G 1.103 0.8363 1.326 98 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.926 0.903 0.924 0.66 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 60/80 7.65G 1.094 0.8231 1.318 133 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.927 0.903 0.924 0.66 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 61/80 7.65G 1.088 0.8159 1.312 111 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.927 0.904 0.925 0.661 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 62/80 7.65G 1.088 0.8158 1.314 95 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.927 0.904 0.924 0.661 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 63/80 7.65G 1.086 0.8115 1.312 114 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.928 0.905 0.926 0.661 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 64/80 7.65G 1.08 0.8129 1.312 90 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.928 0.905 0.926 0.662 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 65/80 7.65G 1.082 0.8094 1.313 85 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.927 0.906 0.926 0.662 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 66/80 7.65G 1.077 0.8019 1.303 125 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.929 0.906 0.927 0.662 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 67/80 7.65G 1.072 0.789 1.298 123 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.929 0.906 0.927 0.662 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 68/80 7.65G 1.069 0.792 1.298 110 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.931 0.904 0.927 0.663 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 69/80 7.65G 1.063 0.7877 1.298 94 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.931 0.905 0.927 0.663 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 70/80 7.65G 1.057 0.7794 1.295 99 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.932 0.907 0.928 0.664 Closing dataloader mosaic albumentations: Blur(p=0.01, blur_limit=(3, 7)), MedianBlur(p=0.01, blur_limit=(3, 7)), ToGray(p=0.01, num_output_channels=3, method='weighted_average'), CLAHE(p=0.01, clip_limit=(1.0, 4.0), tile_grid_size=(8, 8)) Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 71/80 7.65G 0.9323 0.5644 1.212 48 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.931 0.908 0.928 0.664 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 72/80 7.65G 0.917 0.5446 1.198 36 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.931 0.908 0.928 0.664 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 73/80 7.65G 0.9057 0.5325 1.189 30 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.931 0.909 0.929 0.664 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 74/80 7.65G 0.9025 0.5259 1.192 42 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.931 0.909 0.928 0.665 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 75/80 7.65G 0.8977 0.5201 1.189 38 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.93 0.91 0.928 0.665 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 76/80 7.65G 0.8858 0.5093 1.177 42 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.932 0.909 0.928 0.666 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 77/80 7.65G 0.8829 0.5046 1.18 41 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.932 0.908 0.929 0.666 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 78/80 7.65G 0.8735 0.5023 1.175 46 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.933 0.909 0.929 0.666 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 79/80 7.65G 0.8693 0.496 1.167 43 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.933 0.909 0.929 0.667 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 80/80 7.65G 0.8675 0.4929 1.166 34 640: 1 Class Images Instances Box(P R mAP50 m all 3718 5274 0.932 0.909 0.929 0.667 80 epochs completed in 8.459 hours. Optimizer stripped from runs/detect/train/weights/last.pt, 22.5MB Optimizer stripped from runs/detect/train/weights/best.pt, 22.5MB Validating runs/detect/train/weights/best.pt... Ultralytics 8.3.126 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB) Model summary (fused): 72 layers, 11,126,358 parameters, 0 gradients, 28.4 GFLOPs Class Images Instances Box(P R mAP50 m all 3718 5274 0.927 0.902 0.928 0.665 fire 1859 2708 0.925 0.915 0.941 0.668 smoke 2268 2566 0.929 0.889 0.914 0.662 Speed: 0.1ms preprocess, 9.0ms inference, 0.0ms loss, 0.8ms postprocess per image Results saved to runs/detect/train π‘ Learn more at https://docs.ultralytics.com/modes/train
InΒ [Β ]:
!yolo task=detect mode=predict model="/kaggle/working/runs/detect/train/weights/best.pt" device=0 source="/kaggle/working/combined_dataset_final6/test/images" save=True
InΒ [11]:
from ultralytics import YOLO
model = YOLO("/kaggle/working/runs/detect/train/weights/best.pt")
# Evaluate on Training Data
train_results = model.val(data="/kaggle/working/combined_dataset_final6/data.yaml", split="train")
print(f"Train mAP@50: {train_results.box.map50:.4f}")
print(f"Train mAP@50-95: {train_results.box.map:.4f}")
Ultralytics 8.3.126 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB)
Model summary (fused): 72 layers, 11,126,358 parameters, 0 gradients, 28.4 GFLOPs
val: Fast image access β
(ping: 0.0Β±0.0 ms, read: 747.7Β±473.2 MB/s, size: 27.0 KB)
val: Scanning /kaggle/working/combined_dataset_final6/train/labels.cache... 21725 images, 1204 backgrounds, 0 corrupt: 100%|ββββββββββ| 21725/21725 [00:00<?, ?it/s]
WARNING β οΈ Box and segment counts should be equal, but got len(segments) = 22, len(boxes) = 31534. To resolve this only boxes will be used and all segments will be removed. To avoid this please supply either a detect or segment dataset, not a detect-segment mixed dataset.
Class Images Instances Box(P R mAP50 mAP50-95): 100%|ββββββββββ| 1358/1358 [02:45<00:00, 8.19it/s]
all 21725 31534 0.93 0.868 0.934 0.689
fire 10026 16051 0.909 0.85 0.918 0.641
smoke 13622 15483 0.951 0.887 0.95 0.737
Speed: 0.2ms preprocess, 4.3ms inference, 0.0ms loss, 0.9ms postprocess per image
Results saved to runs/detect/val3
Train mAP@50: 0.9336
Train mAP@50-95: 0.6891
InΒ [13]:
# Evaluate on Test Data
test_results = model.val(data="/kaggle/working/combined_dataset_final6/data.yaml", split="test")
print(f"Test mAP@50: {test_results.box.map50:.4f}")
print(f"Test mAP@50-95: {test_results.box.map:.4f}")
Ultralytics 8.3.126 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB)
val: Fast image access β
(ping: 0.0Β±0.0 ms, read: 522.9Β±333.5 MB/s, size: 25.5 KB)
val: Scanning /kaggle/working/smokeandfire-2/test/labels... 543 images, 1 backgrounds, 0 corrupt: 100%|ββββββββββ| 543/543 [00:00<00:00, 1208.44it/s]
val: New cache created: /kaggle/working/smokeandfire-2/test/labels.cache
Class Images Instances Box(P R mAP50 mAP50-95): 100%|ββββββββββ| 34/34 [00:05<00:00, 6.35it/s]
all 543 1065 0.942 0.883 0.896 0.775
fire 528 748 0.959 0.959 0.962 0.807
smoke 190 317 0.926 0.808 0.83 0.744
Speed: 0.6ms preprocess, 4.4ms inference, 0.0ms loss, 1.3ms postprocess per image
Results saved to runs/detect/val5
Test mAP@50: 0.8961
Test mAP@50-95: 0.7755
InΒ [14]:
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
img = mpimg.imread('/kaggle/working/runs/detect/train/results.png')
imgplot = plt.imshow(img)
plt.show()
InΒ [24]:
img = mpimg.imread('/kaggle/working/runs/detect/train/P_curve.png')
imgplot = plt.imshow(img)
plt.show()
InΒ [16]:
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
img = mpimg.imread('/kaggle/working/runs/detect/train/labels.jpg')
imgplot = plt.imshow(img)
plt.show()
InΒ [17]:
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
img = mpimg.imread('/kaggle/working/runs/detect/train/confusion_matrix.png')
imgplot = plt.imshow(img)
plt.show()
InΒ [18]:
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
img = mpimg.imread('/kaggle/working/runs/detect/train/train_batch0.jpg')
imgplot = plt.imshow(img)
plt.show()
InΒ [19]:
from glob import glob
from PIL import Image
import numpy as np
img_path = glob('/kaggle/working/runs/detect/predict/*.jpg')
img_path = np.random.choice(img_path, 10)
for image_path in img_path:
plt.imshow(Image.open(image_path))
plt.axis("off")
plt.show()
print("\n")
InΒ [20]:
!yolo predict model="/kaggle/working/runs/detect/train/weights/best.pt", source="/kaggle/input/fireeee/fire.109.jpg"
WARNING β οΈ argument 'model=/kaggle/working/runs/detect/train/weights/best.pt,' does not require trailing comma ',', updating to 'model=/kaggle/working/runs/detect/train/weights/best.pt'.
Ultralytics 8.3.126 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB)
Model summary (fused): 72 layers, 11,126,358 parameters, 0 gradients, 28.4 GFLOPs
image 1/1 /kaggle/input/fireeee/fire.109.jpg: 384x640 2 fires, 51.6ms
Speed: 4.9ms preprocess, 51.6ms inference, 168.1ms postprocess per image at shape (1, 3, 384, 640)
Results saved to runs/detect/predict2
π‘ Learn more at https://docs.ultralytics.com/modes/predict
InΒ [21]:
output_image_path = "/kaggle/working/runs/detect/predict2/fire.109.jpg"
output_image = cv2.imread(output_image_path)
output_image = cv2.cvtColor(output_image, cv2.COLOR_BGR2RGB)
plt.figure(figsize=(8, 8))
plt.imshow(output_image)
plt.title('Detected Image - fire.109.jpg')
plt.axis('off')
plt.show()
InΒ [22]:
!yolo predict model="/kaggle/working/runs/detect/train/weights/best.pt", source="/kaggle/input/dataaa88/download (4).jpeg"
WARNING β οΈ argument 'model=/kaggle/working/runs/detect/train/weights/best.pt,' does not require trailing comma ',', updating to 'model=/kaggle/working/runs/detect/train/weights/best.pt'.
Ultralytics 8.3.126 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB)
Model summary (fused): 72 layers, 11,126,358 parameters, 0 gradients, 28.4 GFLOPs
image 1/1 /kaggle/input/dataaa88/download (4).jpeg: 416x640 1 smoke, 39.0ms
Speed: 6.3ms preprocess, 39.0ms inference, 155.5ms postprocess per image at shape (1, 3, 416, 640)
Results saved to runs/detect/predict3
π‘ Learn more at https://docs.ultralytics.com/modes/predict
doneΒΆ
try to change layer of yolo by add c2f layerΒΆ
InΒ [3]:
!yolo train model='/kaggle/input/yoloyw/other/default/1/yolov8.yaml' data='/kaggle/working/smokeandfire-2/data.yaml' epochs=120 imgsz=640 batch=32 augment=True lr0=0.005 lrf=0.01 weight_decay=0.0005 optimizer=SGD mosaic=True hsv_h=0.015 hsv_s=0.7 hsv_v=0.4 degrees=15 translate=0.1 scale=0.5 shear=0.2 flipud=0.5 fliplr=0.5 mixup=0.2
WARNING β οΈ no model scale passed. Assuming scale='n'. Ultralytics 8.3.130 π Python-3.10.14 torch-2.4.0 CUDA:0 (Tesla P100-PCIE-16GB, 16269MiB) engine/trainer: agnostic_nms=False, amp=True, augment=True, auto_augment=randaugment, batch=32, bgr=0.0, box=7.5, cache=False, cfg=None, classes=None, close_mosaic=10, cls=0.5, conf=None, copy_paste=0.0, copy_paste_mode=flip, cos_lr=False, cutmix=0.0, data=/kaggle/working/smokeandfire-2/data.yaml, degrees=15, deterministic=True, device=None, dfl=1.5, dnn=False, dropout=0.0, dynamic=False, embed=None, epochs=120, erasing=0.4, exist_ok=False, fliplr=0.5, flipud=0.5, format=torchscript, fraction=1.0, freeze=None, half=False, hsv_h=0.015, hsv_s=0.7, hsv_v=0.4, imgsz=640, int8=False, iou=0.7, keras=False, kobj=1.0, line_width=None, lr0=0.005, lrf=0.01, mask_ratio=4, max_det=300, mixup=0.2, mode=train, model=/kaggle/input/yoloyw/other/default/1/yolov8.yaml, momentum=0.937, mosaic=True, multi_scale=False, name=train, nbs=64, nms=False, opset=None, optimize=False, optimizer=SGD, overlap_mask=True, patience=100, perspective=0.0, plots=True, pose=12.0, pretrained=True, profile=False, project=None, rect=False, resume=False, retina_masks=False, save=True, save_conf=False, save_crop=False, save_dir=runs/detect/train, save_frames=False, save_json=False, save_period=-1, save_txt=False, scale=0.5, seed=0, shear=0.2, show=False, show_boxes=True, show_conf=True, show_labels=True, simplify=True, single_cls=False, source=None, split=val, stream_buffer=False, task=detect, time=None, tracker=botsort.yaml, translate=0.1, val=True, verbose=True, vid_stride=1, visualize=False, warmup_bias_lr=0.1, warmup_epochs=3.0, warmup_momentum=0.8, weight_decay=0.0005, workers=8, workspace=None Downloading https://ultralytics.com/assets/Arial.ttf to '/root/.config/Ultralytics/Arial.ttf'... 100%|ββββββββββββββββββββββββββββββββββββββββ| 755k/755k [00:00<00:00, 39.1MB/s] WARNING β οΈ no model scale passed. Assuming scale='n'. from n params module arguments 0 -1 1 464 ultralytics.nn.modules.conv.Conv [3, 16, 3, 2] 1 -1 1 1888 ultralytics.nn.modules.block.C2f [16, 16, 1, True] 2 -1 1 4672 ultralytics.nn.modules.conv.Conv [16, 32, 3, 2] 3 -1 1 7360 ultralytics.nn.modules.block.C2f [32, 32, 1, True] 4 -1 1 18560 ultralytics.nn.modules.conv.Conv [32, 64, 3, 2] 5 -1 2 49664 ultralytics.nn.modules.block.C2f [64, 64, 2, True] 6 -1 1 73984 ultralytics.nn.modules.conv.Conv [64, 128, 3, 2] 7 -1 2 197632 ultralytics.nn.modules.block.C2f [128, 128, 2, True] 8 -1 1 295424 ultralytics.nn.modules.conv.Conv [128, 256, 3, 2] 9 -1 1 460288 ultralytics.nn.modules.block.C2f [256, 256, 1, True] 10 -1 1 164608 ultralytics.nn.modules.block.SPPF [256, 256, 5] 11 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 12 [-1, 6] 1 0 ultralytics.nn.modules.conv.Concat [1] 13 -1 1 148224 ultralytics.nn.modules.block.C2f [384, 128, 1] 14 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest'] 15 [-1, 4] 1 0 ultralytics.nn.modules.conv.Concat [1] 16 -1 1 37248 ultralytics.nn.modules.block.C2f [192, 64, 1] 17 -1 1 36992 ultralytics.nn.modules.conv.Conv [64, 64, 3, 2] 18 [-1, 12] 1 0 ultralytics.nn.modules.conv.Concat [1] 19 -1 1 156416 ultralytics.nn.modules.block.C2f [448, 128, 1] 20 -1 1 147712 ultralytics.nn.modules.conv.Conv [128, 128, 3, 2] 21 [-1, 9] 1 0 ultralytics.nn.modules.conv.Concat [1] 22 -1 1 493056 ultralytics.nn.modules.block.C2f [384, 256, 1] 23 [15, 18, 21] 1 3481942 ultralytics.nn.modules.head.Detect [2, [192, 448, 384]] YOLOv8 summary: 137 layers, 5,776,134 parameters, 5,776,118 gradients, 21.7 GFLOPs Freezing layer 'model.23.dfl.conv.weight' AMP: running Automatic Mixed Precision (AMP) checks... Downloading https://github.com/ultralytics/assets/releases/download/v8.3.0/yolo11n.pt to 'yolo11n.pt'... 100%|βββββββββββββββββββββββββββββββββββββββ| 5.35M/5.35M [00:00<00:00, 176MB/s] AMP: checks passed β train: Fast image access β (ping: 0.0Β±0.0 ms, read: 1454.3Β±594.5 MB/s, size: 53.8 KB) train: Scanning /kaggle/working/smokeandfire-2/train/labels... 8243 images, 11 b train: New cache created: /kaggle/working/smokeandfire-2/train/labels.cache WARNING β οΈ Box and segment counts should be equal, but got len(segments) = 12, len(boxes) = 16365. To resolve this only boxes will be used and all segments will be removed. To avoid this please supply either a detect or segment dataset, not a detect-segment mixed dataset. albumentations: Blur(p=0.01, blur_limit=(3, 7)), MedianBlur(p=0.01, blur_limit=(3, 7)), ToGray(p=0.01, num_output_channels=3, method='weighted_average'), CLAHE(p=0.01, clip_limit=(1.0, 4.0), tile_grid_size=(8, 8)) val: Fast image access β (ping: 0.0Β±0.0 ms, read: 648.9Β±385.1 MB/s, size: 40.7 KB) val: Scanning /kaggle/working/smokeandfire-2/valid/labels... 1062 images, 3 back val: New cache created: /kaggle/working/smokeandfire-2/valid/labels.cache Plotting labels to runs/detect/train/labels.jpg... optimizer: SGD(lr=0.005, momentum=0.937) with parameter groups 61 weight(decay=0.0), 68 weight(decay=0.0005), 67 bias(decay=0.0) Image sizes 640 train, 640 val Using 4 dataloader workers Logging results to runs/detect/train Starting training for 120 epochs... Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 1/120 5.96G 3.36 3.73 3.79 62 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.515 0.061 0.00815 0.00331 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 2/120 5.98G 3.049 3.223 3.348 92 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.284 0.161 0.0903 0.025 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 3/120 6G 2.478 2.922 2.903 72 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.389 0.301 0.24 0.0874 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 4/120 5.99G 2.089 2.575 2.519 90 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.669 0.459 0.502 0.256 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 5/120 6G 1.945 2.37 2.351 69 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.739 0.572 0.627 0.329 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 6/120 5.99G 1.845 2.215 2.246 65 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.74 0.618 0.656 0.382 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 7/120 6.01G 1.798 2.135 2.203 70 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.756 0.628 0.678 0.412 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 8/120 5.98G 1.758 2.095 2.158 71 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.795 0.61 0.653 0.405 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 9/120 6G 1.719 2.03 2.118 69 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.789 0.682 0.719 0.447 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 10/120 5.99G 1.697 2.007 2.092 85 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.764 0.665 0.711 0.454 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 11/120 6G 1.666 1.953 2.054 87 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.792 0.676 0.717 0.463 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 12/120 5.98G 1.642 1.907 2.025 77 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.796 0.705 0.75 0.511 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 13/120 6G 1.627 1.876 2.012 87 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.817 0.698 0.757 0.504 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 14/120 5.98G 1.596 1.834 1.985 70 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.837 0.714 0.762 0.513 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 15/120 6G 1.59 1.829 1.977 90 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.823 0.726 0.766 0.531 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 16/120 5.98G 1.573 1.801 1.959 111 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.828 0.735 0.771 0.542 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 17/120 6G 1.56 1.766 1.946 64 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.838 0.729 0.77 0.528 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 18/120 5.98G 1.549 1.742 1.939 65 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.834 0.734 0.774 0.544 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 19/120 6G 1.542 1.729 1.936 75 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.838 0.744 0.788 0.538 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 20/120 5.98G 1.523 1.699 1.918 92 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.852 0.752 0.794 0.551 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 21/120 6G 1.511 1.685 1.905 91 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.822 0.757 0.798 0.563 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 22/120 5.98G 1.511 1.676 1.906 107 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.839 0.75 0.792 0.558 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 23/120 6G 1.496 1.667 1.904 117 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.841 0.771 0.8 0.573 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 24/120 5.98G 1.502 1.667 1.898 87 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.853 0.758 0.801 0.565 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 25/120 6G 1.474 1.638 1.88 66 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.848 0.76 0.798 0.565 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 26/120 5.98G 1.49 1.649 1.884 88 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.862 0.771 0.804 0.564 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 27/120 6G 1.461 1.613 1.86 77 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.856 0.774 0.811 0.579 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 28/120 5.98G 1.453 1.6 1.849 79 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.87 0.78 0.818 0.577 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 29/120 6G 1.453 1.59 1.847 119 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.852 0.78 0.815 0.579 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 30/120 5.98G 1.46 1.592 1.854 81 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.87 0.793 0.82 0.586 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 31/120 6G 1.441 1.559 1.836 85 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.852 0.784 0.816 0.591 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 32/120 5.98G 1.431 1.551 1.829 91 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.873 0.781 0.815 0.582 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 33/120 6G 1.433 1.556 1.827 78 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.859 0.795 0.829 0.598 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 34/120 5.98G 1.435 1.541 1.828 83 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.868 0.789 0.817 0.589 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 35/120 6G 1.424 1.532 1.821 52 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.874 0.796 0.83 0.604 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 36/120 5.98G 1.407 1.519 1.809 80 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.87 0.797 0.825 0.6 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 37/120 6G 1.421 1.534 1.823 80 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.854 0.799 0.823 0.602 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 38/120 5.98G 1.413 1.524 1.809 82 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.86 0.798 0.828 0.609 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 39/120 6G 1.421 1.53 1.821 73 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.872 0.798 0.828 0.601 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 40/120 5.98G 1.413 1.506 1.808 75 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.886 0.795 0.833 0.609 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 41/120 6G 1.42 1.53 1.815 87 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.862 0.815 0.84 0.606 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 42/120 5.98G 1.397 1.488 1.796 103 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.876 0.812 0.832 0.608 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 43/120 6G 1.386 1.469 1.789 86 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.871 0.805 0.831 0.611 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 44/120 5.98G 1.39 1.46 1.786 76 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.871 0.806 0.83 0.615 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 45/120 6G 1.386 1.459 1.787 89 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.869 0.797 0.828 0.615 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 46/120 5.98G 1.394 1.473 1.788 90 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.888 0.807 0.837 0.616 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 47/120 6G 1.384 1.477 1.788 89 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.885 0.812 0.839 0.615 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 48/120 5.98G 1.384 1.465 1.784 105 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.88 0.807 0.837 0.617 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 49/120 6G 1.372 1.442 1.77 87 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.877 0.813 0.838 0.62 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 50/120 5.98G 1.37 1.431 1.765 81 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.88 0.822 0.839 0.624 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 51/120 6G 1.364 1.433 1.767 116 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.872 0.821 0.842 0.622 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 52/120 5.98G 1.36 1.434 1.762 56 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.875 0.818 0.842 0.627 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 53/120 6G 1.371 1.44 1.769 92 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.884 0.817 0.846 0.626 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 54/120 5.98G 1.366 1.425 1.757 71 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.875 0.825 0.844 0.622 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 55/120 6G 1.341 1.406 1.747 67 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.875 0.832 0.842 0.628 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 56/120 5.98G 1.36 1.423 1.762 99 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.886 0.817 0.846 0.627 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 57/120 6G 1.347 1.409 1.749 81 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.879 0.822 0.845 0.63 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 58/120 5.98G 1.363 1.415 1.761 101 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.885 0.812 0.84 0.628 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 59/120 6G 1.349 1.4 1.748 101 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.879 0.825 0.843 0.628 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 60/120 5.98G 1.351 1.411 1.749 89 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.879 0.818 0.846 0.628 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 61/120 6G 1.349 1.399 1.745 87 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.878 0.821 0.841 0.627 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 62/120 5.98G 1.341 1.38 1.738 77 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.881 0.824 0.844 0.63 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 63/120 6G 1.35 1.399 1.745 68 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.881 0.823 0.843 0.633 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 64/120 5.98G 1.329 1.37 1.727 67 640: 1 Class Images Instances Box(P R mAP50 m all 1062 2081 0.883 0.819 0.844 0.633 Epoch GPU_mem box_loss cls_loss dfl_loss Instances Size 65/120 6G 1.339 1.384 1.739 134 640: ^C 65/120 6G 1.339 1.384 1.739 134 640:
InΒ [Β ]:
# there is no improvement so as it
InΒ [Β ]:
InΒ [Β ]: